Evaluation of Assays for Epoxides in Oxidized Lipids

by

Chen-Hsiang Liao

A thesis submitted to the

Graduate School-New Brunswick

Rutgers, The State University of New Jersey

in partial fulfillment of the requirements

for the degree of

Masters of Science

Graduate Program in Food Science

written under the direction of

Professor Karen M. Schaich

and approved by

______________________________

______________________________

_______________________________

New Brunswick, New Jersey

May, 2013

ii

ABSTRACT OF THE THESIS

Evaluation of Assays for Epoxides in Oxidized Lipids

BY CHEN-HSIANG LIAO

Thesis director:

Dr. Karen M. Schaich

Epoxides have long been recognized as lipid oxidation products, and there is

recent evidence that epoxides may be as or more important than hydroperoxides under

many conditions. Nevertheless, epoxides are seldom analyzed when monitoring

oxidative degradation, at least in part because there are few established analytical

procedures. The goal of this thesis research, therefore, was to evaluate four available

epoxide assays for accuracy, sensitivity, stoichiometry, reproducibility, and handling

requirements, and from the results provide analytical protocols and practical guidelines

for selection and application of assays for lipid epoxides.

AOCS standard HBr titration of epoxides, nitrobenzylpyridine (NBP) reaction

with colorimetric endpoint, diethyl dithiocarbamate (DETC) complexation with high

iii

pressure liquid chromatography (HPLC) separation and quantitation of adducts, and 1H

nuclear magnetic resonance (NMR) analysis of epoxides were evaluated using

epoxybutane, epoxyhexene, and epoxydecene as standards.

The HBr assay is too insensitive (detects 0.0075-0.1M) for following lipid

epoxides in foods and biological materials. In addition, the reaction must be run under

inert atmosphere to prevent non-specific oxidation of the Br, and HBr degrades so

rapidly that frequent restandardization is necessary.

Nitrobenzylpyridine assay is more sensitive, detecting 0.5 mM epoxides. However,

the NBP reaction response varied considerably with time and temperature of reaction

and with epoxide structure (increased with epoxide chain length). Hence, a different

standard must be used for each epoxide analyzed, and selection of an appropriate

standard for epoxides of unknown structure is problematic.

For the DETC assay, reaction response was linear from 1M to 1 mM for all three

epoxide standards (R2 > 0.99) and oxidized methyl linoleate (R

2 > 0.94), and increased

with epoxide chain length. HPLC analysis of adducts allows differentiation and

quantitation of individual epoxides, so can provide important information about

oxidation chemistry as well as quantitation.

NMR offers the advantage of direct analysis of oils and extracts, detects

micromolar epoxides, and clearly distinguishes epoxides from other oxidation products

iv

in lipids. Response curves were linear with concentration of epoxide standards,

oxidized corn oil, and oxidized methyl linoleate (R2> 0.98).

The DETC-HPLC and NMR assays hold the greatest promise for routine analysis

of epoxides in oxidized lipids.

v

ACKNOWLEDGEMENTS

I would like to thank my advisor, Dr. Karen M. Schaich, for her guidance and

support during the whole process of my graduate study. Her kindness, knowledge,

generosity and patience gave me great help in the pursuit of my Master’s Degree.

Without her, I would have never been able to achieve what I have done today. I am very

grateful for the privilege of having known and worked with her.

I thank my lab mates, Wan Zunair Wan Ibadullah, Xin Tian, JiaXie, Teng Peng, and

Linhong Yao, for their help with my research and also for their friendship.

Finally, I would like to thank my parents and friends for their unconditional

support and constant encouragement in helping me with this endeavor.

vi

TABLE OF CONTENTS

ABSTRACT OF THE THESIS ii

ACKNOWLEDGEMENTS v

TABLE OF CONTENTS vi

LIST OF TABLE ix

LIST OF FIGURE xi

1. INTRODUCTION 1

2. BACKGROUND 3

2.1 General Overview of Lipid Oxidation 3

2.1.1. Initiation of lipid oxidation 3

2.1.2 Propagation 7

2.1.3 Termination 11

2.1.4 Reaction schemes for lipid oxidation 14

2.1.4.1 Classical radical chain reaction 14

2.1.4.2 Integrated alternated pathways of lipid oxidation 14

2.1.5 Factors that affect oxidation 18

2.2 Chemistry and Characteristics of Epoxides 22

2.2.1 Structure of epoxides 22

2.2.2 Formation of epoxides 23

2.2.3 Reactions of epoxides 24

vii

2.2.4 Toxicity of epoxides 25

2.3 Assays for Lipid Epoxides 26

2.3.1 General Considerations 26

2.3.2 HBr (Hydrobromic acid) Assay 28

2.3.3 NBP (4-p-nitrobenzyl-pyridine) Assay 29

2.3.4 DETC (N,N-Diethyldithiocarbamate) Assay with HPLC separation and

quantification of epoxy adducts 31

2.5.5 NMR assay for epoxides 33

3. EXPERIMENTAL PROCEDURES 36

3.1 Overall experimental design 36

3.2 General handling issues 37

3.3 HBr (Hydrobromic acid) Assay(AOCS Cd9-57) Methods and Material 37

3.4 NBP (4-p-nitrobenzyl-pyridine) Assay Methods and Materials 42

3.5 HPLC-DETC (N,N-Diethyldithiocarbamate) Assay

Materials and Methods 47

3.6 NMR (Nuclear Magnetic Resonance) Assay Materials and Methods 51

4 . RESULTS AND DISCUSSION 54

4.1 HBr (Hydrobromic acid) Assay 54

4.2 NBP (4-p-nitrobenzyl-pyridine) Assay 63

viii

4.3 HPLC-DETC (N,N-diethyldithiocarbamate) Assay 80

4.4 NMR (Nuclear Magnetic Resonance) Assay 91

5 SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS 105

6 FUTURE WORK 111

7.REFERENCES 114

CURRICUM VITA…………………………………………………………..123

ix

LIST OF TABLES

Table 1. Activation energy of different propagation reactions 5

Table 2. Energies at various light wavelength compared to bond

dissociation energies for typical chemical bonds 7

Table 3. Lifetimes and hydrogen abstraction rates of various radicals 8

Table 4. Percent oxirane oxygen detected in different concentrations of

epoxydecene, epoxyhexene and epoxybutane standards by

HBr titration 55

Table 5. HBr titration detection of epoxides in oxidized corn oil and

methyl linoleate 56

Table 6. Reproducibility of HBr titration assay for C10 1,2-epoxy-9-decene 57

Table 7. Reproducibility of HBr titration assay for C6 1,2-epoxy-5-hexene 58

Table 8. Reproducibility of HBr titration assay for C4 2,3-epoxbutane 60

Table 9. Reproducibility of HBr titration assay for oxidized methyl linoleate 61

Table 10. Reproducibility of HBr titration assay for oxidized corn oil 60

Table 11. Coefficients of variation for HBr titration of standard epoxides 61

Table 12. Active detection ranges for NBP assay 67

Table 13. Coefficients of variation for the NBP detection of epoxides

in standards and oxidized lipids 74

x

Table 14. Reproducibility of NBP assay for 1,2-epoxy-9-decene standard 75

Table 15. Reproducibility of NBP assay for 1,2-epoxy-5-hexene standard 76

Table 16. Reproducibility of NBP assay for 2,3-epoxybutane standard 77

Table 17. Reproducibility of NBP assay for oxidized methyl linoleate 78

Table 18. Reproducibility of NBP assay for oxidized corn oil 78

Table 19. Coefficients of variation for HPLC-DETC assay of standard epoxides and

oxidized lipids, averaged over all concentrations and days 86

Table 20. Reproducibility of HPLC-DETC assay for epoxybutane 87

Table 21. Reproducibility of HPLC-DETC assay for C6 1,2-epoxy5-hexene 88

Table 22. Reproducibility of HPLC-DETC assay for C10 1,2-epoxy-9-decene 89

Table 23. Reproducibility of HPLC-DETC assay for oxidized methyl linoleate 89

Table 24. Reproducibility of HPLC-DETC assay for oxidized corn oil 90

Table 25. Reproducibility of NMR assay for C10 1,2-epoxy-9-decene 101

Table 26. Reproducibility of NMR assay for epoxides in oxidized methyl

linoleate 102

Table 27. Reproducibility of NMR assay for epoxides in oxidized corn oil 103

Table 28. Coefficients of variation for NMR assay of epoxydecene and

epoxides in oxidized lipids 104

Table 29. Summary of detection ranges and important characteristics

of HBr, NBP, DETC, and NMR epoxides 107

xi

LIST OF FIGURES

Figure 1. Effect of oxygen and temperature on termination processes

in lipid oxidation 12

Figure 2. Example of an epioxide intermediate derived from

Photosensitization undergoing secondary scissions to form

short chain odor compounds and aldehydes 13

Figure 3. Example of a hydroperoxide undergoing OH- and OOH

-

elimination reaction 13

Figure 4. Free radical chain reactions of lipid oxidation proceed in three stages 15

Figure 5. Integrated theory of lipid oxidation showing alternate reactions

that compete with hydrogen abstraction 17

Figure 6. Structure of epoxides relative to ethers 22

Figure 7. The mechanism of how NBP act as a cucleophile to react with epoxide to

form the NBP-epoxide complex 30

Figure 8. The mechanism of DETC to react with epoxide molecule and

followed by an acid decomposition reaction 32

Figure 9. Changes in 1H NMR spectra of oxidizing methyl oleate over time 34

Figure 10. Schematic diagram of the experimental design used to evaluate epoxide

assays 36

Figure 11. Titration curves for reaction of HBr with epoxybutane, epoxyhexene, and

xii

epoxydecene standards 55

Figure 12. Effects of heating time on NBP reaction response of three

standard epoxides 64

Figure 13. NBP reaction response curves for epoxybutane, epoxyhexene, and

epoxydecene standards 65

Figure 14. NBP reaction response at very low concentrations of epoxyhexene and

epoxydecene standards (top) and epoxybutane (bottom) 69

Figure 15. Log epoxide concentration vs absorbance (reaction response) for NBP assay.

Top: epoxydecene and epoxyhexene. Bottom: epoxybutane 70

Figure 16. NBP response curves for oxidized methyl linoleate and corn oil. Top: Full

concentration range tested. Bottom: lower concentration range 72

Figure 17. Log ML and corn oil concentration vs absorbance (reaction response) for

NBP assay. Top: full concentration range tested. Bottom: Lower

concentration range (<6 M) 73

Figure 18. HPLC chromatogram of a 1:1:1 mixture of epoxy butane, hexane, and

decene (top). Bottom: Reaction response is linear for micromolar

concentrations 80

Figure 19. HPLC-DETC reaction response curves (reported as peak areas) for

epoxybutane, epoxyhexene, and epoxydecene standards 82

xiii

Figure 20. HPLC-DETC response curves for oxidized methyl linoleate

and corn oil 84

Figure 21. 1H and

13C NMR spectra of 1,2-epoxy-9-decene.

A and B: this study. C and D. Decene spectra fromSpectral

Database for Organic Compounds 93

Figure 22. 1H NMR peak area as a function of log epoxide concentration for

epoxydecene 94

Figure 23. 1H NMR peak area as a function of log epoxide concentration for

epoxydecene 95

Figure 24. 1H NMR spectra of oxidized methyl linoleate (top) and

corn oil (bottom) 97

Figure 25. 1H NMR peak area as a function of oil concentration for oxidized methyl

linoleate and corn oil 98

Figure 26. 1H NMR peak area as a function of log[oil] for oxidized

methyl linoleate and corn oil 99

1

1. INTRODUCTION

Lipid oxidation is the major chemical reaction degrading food quality during food

storage. During food processing and storage, lipids oxidize in the presence of oxygen,

light, micro-organisms, and enzymes (Kan 2006), generating characteristic rancid

off-flavors and odors, browning, and toxic products directly (Schaich 2005). In addition,

all intermediates and products of lipid oxidation attack other molecules, leading to

texture and functional changes, loss of natural colors, and loss of nutritional value

(Schaich 1980; Kan 2006; Schaich 2008). Thus, the ability to accurately analyze the

extent of lipid oxidation in foods is very important for both industrial quality control

and research.

Tracking lipid oxidation is not an easy task because the reactions are complex and

form many different products. In addition, these products are not stable, but decompose

and transform over time. Traditionally, hydroperoxides have been the major product

analyzed. Volatile breakdown products, particularly hexanal, are also commonly

measured. However, these two products alone account for only some oxidation

pathways and do not accurately portray the extent of lipid oxidation in any material.

Current challenges of stabilizing foods reformulated with highly unsaturated oils for

health as well as increased attention to presence of potentially toxic compounds now

demand more complete accounting of lipid oxidation products.

2

Epoxides have long been recognized as lipid oxidation products yet they are

seldom systematically analyzed when monitoring oxidation in any materials. One

reason may be that epoxides are highly unstable due to high ring strain, and they are

highly reactive with proteins and other compounds (Schaich 2008). Therefore, they

they may not accumulate to detectable levels. A second reason is lack of sensitive

methods for detecting epoxides. Whatever the reason, several studies have shown that

epoxides are the dominant, if not only, product in aprotic systems (Gardner et al, 1978;

Gardner and Kleiman, 1981; Haynes and Vonwiller, 1990) and recent research in our

laboratory has verified epoxides as major products that parallel or exceed

hydroperoxides in concentration. Consequently, it is mandatory that epoxides be

accounted for systematically, in addition to hydroperoxides and other products, when

analyzing lipid oxidation in any material.

To support development of standard protocols for analysis of lipid epoxides and

encourage inclusion of epoxides in standard lipid oxidation analyses, this thesis

evaluated four epoxide assays for accuracy, sensitivity and detection ranges,

stoichiometry, effects of structure, and linearity of response, as well as required reaction

conditions and handling procedures. The four assays were: NBP

(4-p-nitrobenzyl-pyridine) assay, HBr (Hydrobromic acid) Titration assay,

HPLC-DETC (N,N-diethyldithiocarbamate) assay and NMR assay. Each assay has

3

different limitations for detecting lipid epoxides, so a secondary goal was to determine

appropriate conditions for use of each assay. It is hoped that the information provided in

this thesis will arouse interest in epoxide assays and provide a starting point for more

extensive accounting of the role of epoxides in lipid oxidation.

2. BACKGROUND

2.1 General Overview of Lipid Oxidation

Lipid oxidation has huge negative impacts on food systems, including reduced

shelf life; production of off-flavored aldehydes, ketones and all other secondary

compounds; browning; and co-oxidation of other food molecules (Claxson et al. 1994;

Kamal-Eldin et al. 1997; Frankel 1998; Schaich 2008). Lipid oxidation is a constantly

changing process that is generally considered to occur in three stages -- initiation,

propagation and termination (Schaich 2005).

2.1.1. Initiation of lipid oxidation

Lipid oxidation is not a thermodynamically spontaneous reaction. Ground state

triplet oxygen has two unpaired electrons (O-O

) with parallel spins that cannot

directly add to double bonds of unsaturated fatty acids (Frankel 1998), which are in

singlet state. Thus, initiators are required to start the reaction. During initiation,

initiators such as light, heat, metals, or other radicals provide the high energy needed to

4

remove hydrogen atoms from lipid molecules and form ab initio lipid radicals (L●)

(Schaich 2005):

LH L + H

+ (1)

Metals: Transition metals are perhaps the most active initiator of lipid oxidation in

foods since metals are ubiquitous components and contaminants. Even at

concentrations as low as 0.1 ppm, transition metals possessing two or more valency

states can act as catalysts to initiate lipid oxidation, (Kan 2006). Higher valence state

metals act primarily by electron transfer to cause formation of radicals in lipids, while

reduced metals act primarily by reduction of the O-O bond in hydroperoxides (Schaich,

1992). These reactions will be discussed in more detail in Section 2.2.4.

Heat: Thermal energy at high temperatures (e.g. frying or baking) in the presence

of oxygen induces bond scissions to form radicals (L, LO

and LOO

) which can then

abstract hydrogens to start the radical chains of oxidation (Nawar, 1969, 1986). The

main effect of moderate heat, i.e. under most conditions except frying, is to break O-O

bonds of ROOH or LOOH already formed by enzymes, metals, photosensitizers, or

oxidation, since this is the only propagation step that has any significant activation

energy (Table 1) (Labuza 1971). The RO, LO

, and HO

● free radicals thus generated

then abstract nearby lipids for hydrogens to form L● and new radical chain reactions.

This greatly accelerates the rate of oxidation.

initiator

5

Table 1. Activation energy of different propagation reactions. Data from (Labuza

1971).

ko, oxygenation; kp, propagation; kt, termination; kd, decomposition

Light: Light initiates lipid oxidation less efficiently than heat and metals. Shorter

wavelengths of ultraviolet light theoretically have enough energy to break C-C or C-H

bonds in acyl chains (Table 2), but in practice ionization is more common. Thus, it is

difficult to produce ab initio L radicals directly by light. The main effects of light are in

two other directions (Schaich 2005):

a) Decomposition of low energy (157 kJ/mol) O-O bonds in hydroperoxides by

ultraviolet light. This generates LO and

OH radicals that abstract hydrogens

more rapidly than LOO and accelerate lipid oxidation greatly.

b) Generation of free radicals or singlet oxygen, 1O2, by visible light and

photosensitizers. When exposed to light, photosensitizers absorb energy and

6

jump to an excited state. This energy is released in direct electron transfer to

form free radicals, or in electron transfer to oxygen. The energy transforms

ground state oxygen 3O2 (unpaired electrons in parallel spin) to excited state

oxygen 1O2 (unpaired electrons in opposite spin). Now having the same spin

state as double bonds, 1

O2 can add across lipid double bonds to form

hydroperoxides without generating radicals (Schaich 2005). Due to the high

energy of excited state oxygen 1O2, the reaction rate of singlet oxygen is 1500

times greater than autoxidation (Kan 2006).

Photosensitizers in foods and biological materials are usually pigments such as

chlorophyll and hemoglobin. Type 1 photosensitization (free radical) is oxygen

dependent and is distinguished from normal autoxidation only by accelerated kinetics.

Type 2 photosensitization (singlet oxygen) is oxygen independent and related mostly to

the concentration of sensitizer present (Schaich 2005)

7

Table 2. Energies at various light wavelength compared to bond dissociation energies

for typical chemical bonds (Schaich 2005).

2.1.2 Propagation

During propagation, ab initio lipid radicals react with oxygen to form peroxyl

radicals (LOO). L

radicals are relatively unreactive, but peroxyl radicals abstract

hydrogens from nearby lipid molecules to form a stable intermediate hydroperoxide

(LOOH) and generate a new lipid radical (L’) at the same time (Schaich 2005).

Hydroperoxides are decomposed by light, heats and metals to alkyl radicals (LO)

which are several orders of magnitude greater in reaction rate than peroxyl radicals

(LOO). In the propagation stage, hydroperoxide decomposition is the key rate-limiting

step (Kamal-Eldin et al. 1997).

8

Table 3. Lifetimes and hydrogen abstraction rates of various radicals (Schaich 2005).

The followings are some major competing propagation mechanisms of lipid

peroxyl radicals (Schaich, 2005):

1. Hydrogen abstraction to form hydroperoxides:

LOO + LH LOOH + L

LOO + L’OOH LOOH + L’OO

Peroxyl radicals formed from initiation reactions abstract hydrogen from nearby lipids

or hydroperoxides to form hydroperoxides and new radical molecules (L and L’OO

,

respectively); the new radicals add oxygen and start new chains. This cycle of reactions

can continue indefinitely until the chain is intercepted.

2. Rearrangement/cyclization reaction to form epoxides, hydroxyepoxides and

epidioxides. When abstractrable hydrogens are not immediately available, peroxyl

radicals add to nearby double bonds to form cyclic products. The most important

internal rearrangement of peroxyl radical proceeds by 1,3-addition of the peroxyl

radical to the nearby cis-double bonds to form a 5-exo ring (Rx.2). Initial

9

cyclization of peroxyl radical via 1,4-addition to form a 6-exo exocyclic peroxides

(Rx.3) forms only in fatty acids with four or more double bonds.

(2)

(3)

3. Additions to double bonds to form dimer, polymers and hydropeoxy epoxides

(Kochi 1962,1973). Addition reactions compete with hydrogen abstraction. Although

perhaps less favored under general conditions, addition reactions become competitive

when the abstractable hydrogens are limited (aprotic solvent, low temperature). In early

stages of oxidation, LOO add to double bonds to form initial dimer complexes (Rx. 4)

which can further react to form new radicals. LOO add to isolated or nonconjugated

double bonds, then undergo 1,3-cyclization to generate an epoxide and new radical that

adds oxygen to produce a new peroxyl radical (Rx. 4).

10

LO

(4)

(5)

Alkoxyl radicals also have competing reactions that can change the course of lipid

oxidation (Schaich 2005). Of particular importance to propagation are:

a) internal rearrangements to form epoxides and allylic radicals (Rx. 6) (Kochi

1962, Wu et al. 1977, Haynes and Vonwiller 1990). The radicals add oxygen to

become peroxy(epoxy) radicals, undergo additional reactions, and eventually

decompose to as yet unidentified products.

–HCH=CH–CH=CH–CH-R2 –HCH=CH–CH–CH–CH-R2 (6)

b) addition to double bonds (Rx. 7) (Kochi 1962,1973). This occurs most readily

with conjugated double bonds, so increases in importance as oxidation

progresses. Unlike peroxyl radicals that add almost exclusively to trans double

bonds, alkoxyl radicals preferentially add to cis double bonds (Kochi 1962).

LO

+ (7)

c) scission of the acyl chains at positions and to the alkoxyl radical to form two

radicals (Rx. 8) (Chan et al. 1976). Some of the radicals convert immediately to

carbonyl compounds (termination products), but the remaining radicals can add

O

O

.

11

oxygen to form secondary peroxyl radicals, or can add to double bonds to

generate more complex radicals. This source of propagating radicals creates

complex product distributions yet is commonly overlooked.

O O O

R1–CH–R2 R +

CH–R2 OR R1CH + R2 (8)

2.1.3 Termination

Propagation continues until there are no available hydrogens for abstraction or

until non-radical products are formed by one of the following processes:

1. Radical recombinations of two individual free radicals to form stable

non-radical products. These non-radical materials -- ketones, aldehydes, alcohols,

alkanes and dimers -- are called secondary oxidation products and often result in

flavor change, color change and texture degradation. The types of recombinations

that occur are dependent on reaction conditions, that in turn determine the types of

radicals formed. Oxygen and temperature play particularly important roles in

radical recombination. When oxygen pressure pO2 is low and temperature is high

(rapid thermal scissions active), L recombination dominates. In contrast, when the

oxygen pressure pO2 is high and temperature is low, LOO reactions dominate as

shown in Figure 1.

12

Figure 1. Effect of oxygen and temperature on termination processes in lipid oxidation

(Schaich 2005), redrawn from Labuza (1971).

2. Cleavage of alkoxyl radicals when proton sources are present to stabilize

products. Alkoxyl radicals undergo scission on either side of the –C(O)- to form

aldehydes, alkanes, and other products during lipid oxidation. Any fragments that

contain double bonds can continue to oxidize to form carbonyl, alkane and other

short chain products which contribute to the characteristic odors in lipid oxidation.

Some examples of scission reactions are shown in Figure 2.

3. Co-oxidation of other molecules (radical transfer). Lipid radicals abstract

available hydrogens even from non-lipid molecules such amino acid and protein

(Schaich, 1980; Schaich, 2008; Guillén 2009). This quenches lipid radicals and

ceases the propagation reactions, but transfers the oxidation potential to other types

of molecules, thus broadcasting oxidation damage.

13

Figure 2. Example of an epioxide intermediate derived from photosensitization

undergoing secondary scissions to form short chain odor compounds and

aldehydes (Schaich, 2005; redrawn from Frankel et al., 1982).

4. Eliminations. OH and OOH

can be cleaved from hydroperoxides and form an

internal ketones and paired desaturated products with an additional double bond,

respectively (Figure 3).

Figure 3. Example of a lipid hydroperoxide undergoing OH

and OOH

elimination (Schaich 2005).

14

2.1.4 Reaction schemes for lipid oxidation

2.1.4.1 Classical radical chain reaction

A reaction scheme showing how the free radical chains of lipid oxidation have

been traditionally written is shown in Figure 4. In this series of reactions, hydrogen

abstraction is the only chain propagating process, products do not accumulate until the

end of oxidation, and no mechanisms are specified for the formation of different

products. This scheme is oversimplified, does not account for the multitude of different

products that arise from oxidizing lipids, and is inconsistent with observed kinetics and

appearance of products. Thus, this simplistic reaction sequence must be incomplete.

2.1.4.2 Integrated alternate pathways of lipid oxidation

There is ample evidence in the literature documenting reactivity of peroxyl and

alkoxyl radicals beyond hydrogen abstraction and appearance of other products in

competition with hydroperoxides. These alternate reactions, outlined above, have been

reviewed in detail by Schaich (2005). Alternate reactions of peroxyl radicals, LOO,

that compete with hydrogen abstraction include -scission of oxygen, internal

rearrangement to epidioxides, addition to double bonds, and dismutation. Alternate

competing reactions of alkoxyl radicals, LO, include internal rearrangement to

epoxides, addition to double bonds, and scission to secondary products. How these

alternate reactions may be integrated into a more complete reaction scheme is shown in

15

Free Radical Chain Reactions of Lipid Oxidation

Stage 1: Initiation (starting the chain)

Energy, M, radicals

LH L + H

+

Stage 2 : Propagation (involving additional molecules in the chain)

L + O2 LOO

k > 10

9 M

-1sec

-1 (very fast)

LH + LOO LOOH + L

LOOH LO +

OH

LO

+ LH LOH + L

Stage 3 : Termination

LOO + LOO

L

+ LOO

LO + L

L + L

Figure 4. Free radical chain reactions of lipid oxidation proceed in three stages

Non-radical products

(aldehydes, alkanes, peroxides, epoxides,

ketones, polymers, …etc)

Energy, M

16

Figure 5. The alternate reactions all compete with each other, and the balance between

them changes with reaction conditions. Thus, the standard practice of following lipid

oxidation by hydroperoxides alone, or perhaps in combination with hexanal as a

secondary product, cannot accurately determine the extent of oxidation and may even

be misleading (Claxson et al. 1994). Not only do hydroperoxides decompose, but when

other pathways are active they form at much lower levels, later, or even not at all.

Similarly, when alternate pathways compete with scission of alkoxyl radicals,

hexanal may not be formed. The problem is, favored pathways and types of products

present in materials are seldom known before analysis, so it is impossible to predict

ahead of time what products should be measured. Additional problems arise because

different environmental factors or circumstances promote different lipid oxidation

reaction pathways (Schaich 1992, Jie et al. 2003), and because food systems are

complex matrices and lipid oxidation will be affected and altered by the presence of

other ingredient within food (Marmesat 2008). Unless products from competing

alternate pathways are measured, the wrong pathway can be followed and lipid

oxidation can be missed. This means that accurate tracking of lipid oxidation products

and pathways requires monitoring multiple products simultaneously on each sample

(Schaich 2012).

17

Figure 5. Integrated theory of lipid oxidation showing alternate reactions that compete

with hydrogen abstraction (Schaich, 2005, 2012).

18

2.1.5 Factors that affect oxidation

Oxygen

Lipid oxidation is not a self-initiated process, but requires initiators to form the ab

initio radical that starts the radical chains. This is the rate-limiting step. Heat, light,

metals, lipoxygenase, and other radicals are the most common initiators (Schaich 1980).

Oxygen adds to this radical almost instantaneously, that is, at rates controlled only by

how fast oxygen can diffuse to the radical (Schaich 2005, Schaich 2012). Thus, addition

of oxygen affects oxidation rates only at very low oxygen concentrations. When oxygen

pressure is low (less than 1%), the rate of oxidation can be expressed as

Rate of oxidation (low pO2) = k2(k1/k8)1/2

[LOOH][O] (9)

Under these conditions, the oxidation rate is dependent on oxygen and the balance

between initiation and termination by alkyl radical recombinations (Schaich 2005).

Above some level (few % O2), more than enough oxygen is present to fill all radicals,

so rates becomes independent of oxygen, and oxygen effects shift to product

distributions. Under high oxygen pressure, the rate of oxidation can be expressed as

following equation (Schaich 2005).

Rate of oxidation (high pO2) = k3(k1/k6)1/2

[LOOH][LH] (10)

where K3=H abstraction rate of LO, K1=initiation rate to form L

, and K6=termination

rate by LOOH recombination

19

Metals

Redox active metals are major factors catalyzing lipid oxidation in biological

systems. All redox-active can catalyze lipid oxidation, although metals such Fe, Cu,

and Co are the most common. The simplest mechanism for catalysis of initiation is

direct electron transfer from higher valence state metals to double bonds in lipids to

form lipid alkyl radicals, L (Schaich 1992, 2005):

RCH-C

+HR + M

n+ (11) RCH=CHR + M

(n+1)

R + H

+ + M

n+ [ L

+ RH] (12) RH + M

(n+1)+

RCOO + H

+ + Mn

+ R

+ CO2 (13) RCOOH + M

(n+1)+

RCHO + M(n+1)+

RCO + H

+ + Mn

+ [ L· + scission products ]

Lower valence state (reducing) metals require formation of complexes with oxygen that

generate reactive oxygen radical species in order to initiate radical chains (Schaich,

1992, 2005):

[M(n+1)+

…O2–

] Mn+

+ O2 (14)

(a) M(n+1)+

+ O2–

HOO L

+ H2O2

(b) L + [M

(n+1)+…

–O2H]

(c) L + M

n+ + HOO

L’

+ H2O2

(d) LO + [M

(n+1)+…

–OH]

(e) [M(n+1)+

L–

] + M(n+1)+

+ HOO L’

+ H2O2

LH

LH

L’H

LH

LH

20

Redox-active metals are perhaps even more important in catalyzing propagation

by decomposing relatively unreactive hydroperoxides to very reactive oxyl radicals

(Schaich 1992, 2005):

LOOH + Fe3+

LOO + H

+ + Fe

2+ (15)

LOOH + Fe2+

LO + OH

+ Fe

3+ (16)

However, it must be noted that metals have some very different reaction

mechanisms in aprotic solvents such as lipids, and as a consequence, oxygen insertion

to form ketones is a common, if under-recognized reaction in oils, as is formation of

metal-hydroperoxide complexes can generate hypervalent complexes that yield

epoxides and alcohols in accelerated reactions (Schaich 1992, 2005). Overall, metal

catalysis in multiphasic biological system can become quite complex and involve many

different catalytic mechanisms depending on the phase of solubilization.

Enzymes

Lipoxygenase is an enzyme that specifically catalyzes formation of

hydroperoxides in 1,4-pentadienes, preferably linoleic acid, without a lag period and

without generating radicals. Catalysis rates are affected by several factors, most

importantly pH, pO2, and type and concentration of lipid (Kuo 2002). When

lipoxygenases are active, hydroperoxides are generated at rapid rates to accumulate an

oxidant reservoir. When metals, light or moderate heat are present, decomposition of

21

the hydroperoxides can lead to a cascade of damaging radicals that greatly accelerate

and broaden propagation reactions.

Water

Water activity (Aw) strongly affects the rate and pathways of lipid oxidation.

Oxidation occurs rapidly in very dry foods (Aw = <0.1~0.3), but as moisture is added

and bound, when Aw reaches the monolayer value for the food, the rate of lipid

oxidation drops to a minimum (Labuza 1971, Karel 1980). Reasons given for this

behavior are that in dry systems, the matrix is open and air freely diffuses through it,

metals are bare so are catalytically more active, and hydroperoxides have no

stabilization from hydroigen bonding. As small amounts of water are added up to the

monolayer value, reactive sites are hydrated and oxygen access is inhibited, metals are

hydrated and electron transfer is impeded, and hydroperoxides are stabilized through

hydrogen bonding. However, as water activity increases above the monolayer through

the intermediate moisture region (Aw~ 0.33 to 0.73), the lipid oxidation rate increases

again as catalysts become mobilized and more oxygen becomes dissolved in the water.

Above Aw~0.73, concentrations of catalysts and reactants become diluted by large

amounts of water so lipid oxidation rates decrease (Karel 1980).

22

2.2 Chemistry and Characteristics of Epoxides

2.2.1 Structure of epoxides

One class of lipid oxidation products that has been detected along with other

products but seldom measured in focused analyses is epoxides, the topic of this thesis.

Also known as oxirane and ethylene oxide, epoxides are ethers with the oxygen atom

linked cyclically between the two carbons cyclic, thus forming the corner of a triangle

ring (Clayden 2006).

Epoxides are highly reactive and generally quite volatile. The average bond angle

inside the ring is 60 (Figure 6), in comparison to the ideal angle of 109 for tetrahedral

molecules. This places about 49 of strain on each corner, or approximately 150 of

total strain within each epoxide ring. This strain gives rise to the instability and high

reactivity characteristics of epoxides (Wade 2006).

Figure 6. Structure of epoxides relative to ethers.

Tetrahedral sp3 orbital

Bond angle = 109.5°

Planer sp3 Epoxide Bond angle = 60° Angle strain = 49° (109.5-60) Total strain = ~150° (49*3)

CH CH

O

23

2.2.2 Formation of epoxides

Epoxides are formed generally by addition of neighboring oxygenated compounds

to double bonds. For example, epoxides are formed by addition of the peroxy oxygen of

peroxy acids to double bonds (Clayden 2006):

(17)

In lipids, epoxides form via two main mechanisms – addition of alkoxyl radicals

to adjacent double bonds with formation of an allylic alkyl radical,

R1CH=CH-CH=CH-CHR2 R1CH=CH-CH-CH-CHR2 (18)

and by elimination of an alkoxyl group from a hydroperoxyl adduct, followed by

addition of the remaining oxygen to the adjacent radical:

R1-CH2-CH-CH-R2 + LO (19) R1-CH2-CH-CH-R2

Epoxides can also be formed during heating of oils. Heating of olive and

sunflower oils to 180℃ for 15 hours led to the formation of 1% monoepoxy fatty acids,

largely epoxy oleic acid in olive oil and predominantly linoleic acid in sunflower oil

(Fankhauser-Noti et al. 2006). Direct analysis of thermal oxidized methyl oleate and

linoleate by GC-FID detected methyl trans-9,10-epoxystearate and cis-9,10-epoxy-

stearate as products of methyl oleate while methyl trans-9,10-epoxystearate,

cis-9,10-epoxystearate, methyl trans-12,13-epoxystearate and cis-12,13-epoxystearate

24

were the possible products of methyl linoleate (Giuffrida et al. 2004, Marmesat et al.

2008).

Epoxides are recognized as lipid oxidation products but are seldom measured.

There are several reasons that account for this:

1. Concentrations may be low because epoxides are unstable and rapidly transform

to other products –

Alkenals (Anderson, 1962)

metal-mediated cyclization and internal rearrangement to dihydroxy and

hydroxyene compounds and their isomers (Acott and Beckwith, 1964)

Fe-catalyzed isomerization and conversion of OH-epoxides to ketols

(Tung et al. 1962)

2. Lack of sensitive and accurate assays to detect and quantitate epoxides.

Because epoxides are seldom analyzed systematically but are rather detected

adventitiously in other assays, the actual contribution of epoxides to lipid oxidation is

unknown and most likely grossly underestimated (Schaich 2005).

2.2.3 Reactions of epoxides

The fast reactions of epoxides are important contributors to their instability, their

toxicity, and difficulties in detection. Tracking epoxide formation and reaction is

problematic for several reasons:

25

1. Epoxides are unstable -- after epoxides are formed, they are rapidly degraded to

other substances or products (e.g. alkenals) (Anderson 1962)

2. Metals especially Fe and Cu will catalyze cyclization and internal rearrangement

to form dihydroxy and hydroxyene compounds and their position isomers. (Acott 1964;

Schaich 2005)

3. Fe also catalyzes isomerization and conversion of OH-epoxides to ketols. (Tung

1992)

2.2.4 Toxicity of epoxides.

Epoxides have been reported to be potent mutagens and carcinogens by way of

alkylating nucleic acids (Agarwal et al. 1979, Nelis and Sinsheimer 1981, Bush and

Kozumbo 1983) . Aliphatic epoxides are intermediates in bio-transformation of several

ethylenic compounds. Epoxy compound in vivo can degrade human transmission

molecules. The epoxy-compound 4,5-epoxy-2(E)-decenal is a precursor of

etheno-2-deoxyadenosine, a highly mutagenic substance that adds to human DNA (Lee

et al. 2001, Chen et al. 1998. Epoxides are also involved in a wide range of biological

effects including binding to cellular macro molecules (Agarwal 1979, Nelis and

Sinsheimer 1981, Bush and Kozumbo 1983). Lysine and histidine are the two most

susceptible amino acids in proteins that will be attacked by lipid epoxides (Lederer

1996).

26

2.3 Assays for Lipid Epoxides

2.3.1 General Considerations

Epoxides can be formed during lipid oxidation and they are also intermediate

products during many chemical reactions and even in biochemistry pathways

(Hammock 1974). Epoxide formation is one of the alternate lipid oxidation pathways

that compete with hydrogen abstraction and each other for directing oxidative

breakdown of lipids (Schaich 2005). Since the lipid oxidation reaction pathway is

complicated with many products, both class and individual product analyses are needed

to detect, quantitate, and identify epoxides in oxidizing lipids. Class analyses provide

general mass balances of products from different pathways shown in Figure 5, and this

is useful for tracking balance and direction of oxidation under different conditions.

However, more detailed separations of products with mass spectrometry detection are

needed to identify individual products and determine specific reaction mechanisms

(Ehsan 2010, Guo et al. 2010).

One major reason why epoxides are seldom measured in lipid oxidation is the lack

of sensitive, useful assays for epoxides. To fill this void, this thesis focuses on detecting

epoxides as a whole, evaluating three chemical assays and one physical assay. Key

issues will be sensitivity and accuracy of the chemistry, ease of handling, and

conditions required for accuracy and reproducibility.

27

Even cursory evaluations of epoxide assay methods reported in the literature

reveal diverse working conditions and detection ranges reported for each assay.

Similarly, many precautions and handling issues must be addressed to ensure

generating correct data. Each assay has both pros and cons. Some assays are easily

performed but the detection range is too narrower or too high; some assays may be

accurate but require expensive reagents or instrumentation not readily available in

every lab. These issues should be considered carefully when selecting the assay most

suitable for each application.

Because epoxides are highly reactive, the most common approach for class assays

is to complex the epoxides to a target molecule that will give a detectable signal, or to

convert the epoxides to a stable derivative that can be detected (Kim 1992). In this

thesis, both chemical and physical assays were evaluated. Overall, chemical assays

required more handling, and the resulting error or bias must be accounted for in

calculations and interpretation. For example, when epoxides are titrated with

hydrobromic acid as a reagent, judging the end point is a big challenge.

Physical assays such as NMR or HPLC-DETC assays require less pre-treatment and

handling and are probably more accurate, but how can we know that the experimental

results reflect the actual epoxide composition? As a result, both chemical and physical

assays need to be tested and compared.

28

2.3.2 HBr (Hydrobromic acid) Assay

Based on AOCS standard method Cd 9-57 (AOCS 1997), the HBr assay directly

determines the percentage of oxirane oxygen in samples by titration of the epoxide with

hydrobromic acid (Husain 1989). The acid catalyzes ring opening by an SN1

mechanism and the nucleophilic bromide ion with additional electrons (Br) adds to the

epoxy ring at the less substituted end, as shown in Rx. 20 (Claydon 2006).

H+ Br + (20)

Since this assay is based on titration, stirring is required. Although little attention

has been given to how the speed of stirring influences the reactions, one study showed

that if an epoxy ring is to be formed from the mother material, the speed of stirring is

extremely important. Typically, the reaction rate increases with stirring rate --1000

rpm<1500 rpm< 2000 rpm<2500 rpm (Meshram et al. 2011). Another study showed

that when oxirane contents are less than 0.1%, hydrobromic acid titrations of epoxides

work best if oils are heated to decrease viscosity (Shahidi 1996). However, epoxy

groups degrade very rapidly at elevated temperatures (Earle 1970).

There is some evidence that the presence of cyclopropenoid acids and conjugated

dienols interferes with this reaction (Ansari 1986). Salts of epoxy acids can also react

HC CH

O

H-C

H

H

OH

Br

C-H

29

with amine groups to form amine hydrohalide complexes (Durbetak 1958). Thus, back

titration has been introduced to minimize interference from cyclopropenoid, amine and

conjugated dienols or salts and other molecules that form bonds with hydrobromic acid

(Lee et al. 2009).

2.3.3 NBP (4-p-nitrobenzyl-pyridine) Assay

4-p-Nitrobenzyl-pyridine (NBP) alkylation of epoxides has been used as an assay

for epoxides for about 40 years. Briefly, the epoxide reacts with NBP to form a stable

complex that can be detected and quantitated by its violet color and optical absorbance

at 600 nm (Hammock 1974, Hemminki 1979, Chen et al. 1998). The detection

wavelength required may vary with the structure of the epoxide (Agarwal 1979). The

overall reaction between epoxides and 4-p-nitrobenzyl-pyridine is illustrated in Figure

7. The epoxy ring is attacked by the NBP reagent in an SN2 reaction and the ring is

opened. Epoxides will react with amine group to give amine-alcohol products (Clayden

2006). Heat is generally required for this reaction to occur, but the temperature must be

maintained at less than 180 C to avoid degrading the epoxides (Agarwal 1979).

30

Figure 7. The mechanism of how NBP act as a cucleophile to react with epoxide to form

the NBP-epoxide complex (Kim and Thomas 1992).

This assay has a number of limitations in specificity. Epoxide structure affects

reaction of epoxides with 4-p-nitrobenzyl-pyridine in both pattern and extent of

reaction (Thomas et al. 1992). In addition, 4-p-nitrobenzyl-pyridine alkylates a wide

range of materials in addition to epoxides, including halogenated hydrocarbons,

hydrazine derivatives, aldehydes, thiuram and dithiocarba- mate derivatives

(Hemminki et al.1980, Nelis 1982). Assay results will be inaccurate if those molecules

are present in the sample.

The NBP assay has been used to detect epoxides in vivo as well as in vitro

(Hemminki 1979). As mentioned previously, epoxides are potential mutagens and

arcinogenic reagents, so being able to detect formation early is a great advantage.

Epoxide

NBP Epoxide-NBP complex

31

2.3.4 DETC (N,N-Diethyldithiocarbamate) Assay with HPLC separation and

quantification of epoxy adducts

This assay reacts DETC (N,N-diethyldithiocarbamate) with epoxides to form

stable epoxide-DETC complexes (Benson et al. 1985) which are then separated by

HPLC (high performance liquid chromatography) and detected at 278 nm

(Dupard-Julien et al. 2007). Individual adducts can be identified by comparison with

authentic adducts or by interfacing HPLC with mass spectrometry detection.

This assay requires excess DETC to react with epoxide to ensure all the epoxide

binds to DETC. Two products are then formed -- a major form and a minor form (Figure

8). To eliminate interference with detection, phosphoric acid must be added to decrease

the pH and decompose the remaining DETC in the solution (Dupard-Julien et al. 2007).

Initial applications of this method used normal phase HPLC to analyze epoxide

complexes (Munger et al. 1977, Van Damme et al. 1995). However, normal phase

columns have some critical drawbacks. A normal phase column must be eluted by a

non-polar, non-aqueous mobile phase solvent such as chloroform. However,

chloroform is quite toxic and thus subject to severe restrictions on use. Normal phase

columns also exhibit poor reproducibility in retention time due to the presence of water

and protic organic layer on the column surface. Because of these shortcomings, normal

phase columns lost favor after reversed phase columns were invented (Hemström et al.

32

2006), and current HPLC-DETC assay applications have been modified for reversed

phase columns (Dupard-Julien et al. 2007).

Figure 8. The mechanism of DETC to react with epoxide molecule and followed by an

acid decomposition reaction (Dupard-Julien et al. 2007).

The assay can be used to determine epoxide compounds in cells and tissues

(Rather et al. 2010). However, care is required in this application since DETC is a

reasonably strong metal chelator for metals such as lead (Pb), mercury (Hg), copper

(Cu) and cadmium (Cd) (Van Damme 1995). Thus, metals can be interferences with the

DETC assay, altering the efficiency and stoichiometry of the reaction.

2.3.5 NMR assay for epoxides

NMR itself as a tool is highly expensive and the background knowledge behind

33

NMR is very complicated. Nevertheless, NMR is unmatched as an analytical method

for analysis of molecular structure (Polozov et al. 1986, Baumann et al. 2002). NMR

can provide detailed information about the types and amounts of different functional

groups in a sample (e.g. edible oil) and these can be organized into an overall picture of

the structure and configuration of the (Guillén et al. 2004, Guillén 2006, Guillén 2009).

This includes the ability to accurately distinguish between (Z,E) and (E,E) isomers

( Guillén 2009).

Studies of edible oils by NMR have shown that there is a specific chemical shift

range for the epoxy ring. For 1H NMR the chemical shift range for the epoxy hydrogen

is 2.5~3.5 ppm and for 13

C NMR the chemical shift range for epoxy carbon is 45~55

ppm (John et al. 2003, Ergozhin et al. 2004). Figure 9 shows the epoxy region near 2.9

ppm (–CHOCH–) in 1H NMR of methyl oleate. This clearly increases as the peak at

2.01 ppm corresponding to (–CH2–CH=CH–CH2–) which gives rise to the epoxy

group disappears with incubation time (Aerts and Jacobs 2004). A similar phenomenon

occurs in a) palm oil where epoxide peaks at 2.9-3.1 ppm grow in with oxidation and

disappear when the epoxide group is cleaved (Guo et al. 2010), and b) oxidized epoxy

palm oil in biodegradable lubricants ( Moreno et al. 1999, Piazza et al. 2003).

34

Figure 9. Changes in 1H NMR spectra of oxidizing methyl oleate over time.

Double bond peaks at 2.01 ppm decrease as epoxide peaks at 2.9 ppm increase

(Aerts and Jacobs 2004).

NMR analyses of epoxides offer several advantages. First, NMR spectra provide

information about the total structure of the compound or material – not just epoxides.

In terms of epoxides, NMR can distinguish epoxy stereo isomers. This can be seen in

spectra of 6,7-epoxystearic acid where cis forms have shifts of 2.25 ppm while trans

forms of epoxy fatty acids have chemical shifts at 2.52 ppm (Kannan 1974).

H3C (H2C)6 H2C CH CH CH2 (CH2)2 CH2 COOH

O 6,7-epoxystearic acid

NMR also has the potential to detect other secondary oxidation products in the

2.9 ppm 2.01 ppm

Epoxides Double bonds

35

same spectra as epoxides. This is possible because the chemical shifts for oxygenated

compounds are significantly separated from alkyl functional groups. As one example,

high proportions of cytotoxic and genotoxic 4-hydroperoxy-, 4-hydroxy- and

4,5-epoxy-trans-2-alkenals were formed and recognized during oxidation of sesame oil

at 70℃ (Guillen et al. 2005, Guillén et al. 2006). In addition, chemical shifts and

hyperfine structure of methylic, allylic and bis-allylic hydrogen atoms can provide

information about the degradation of different acyl groups during lipid oxidation

(Guillen et al. 2005).

From a practical standpoint, NMR is time-saving. An NMR spectrum can be

collected in as little as ten minutes -- one cycle (8 scans) (Claxson et al. 1994). Also,

samples require less pre-treatment and handling, so fewer artifacts and extraneous

oxidation are introduced. One issue of concern is eliminating water and other

non-sample related protons from the sample. To eliminate proton interference,

deuterated solvents such as CDCl3 are used as solvents for lipids (Cravero et al. 2000).

36

C4 2,3-epoxybutane,

C6 1,2-epoxy-5-hexene,

C10 1,2-epoxy-9-decene,

oxidized methyl linoleate,

oxidized corn oil

NBP and HBr

R group effect, detection range, handling issues, linear range

HPLC and

NMR

R group effect, detection range, handling issues, linear range

3. EXPERIMENTAL PROCEDURES

3.1 Overall experimental design

Figure 10 shows the overall experimental design for testing four epoxide assays

with three standard epoxides, then applying the assays to analysis of epoxides in

authentic oxidized lipids methyl linoleate and corn oil. Each assay was tested with a

range of standard concentrations to determine linear detection range and detection

limits, R group effects, accuracy of assay, and handling issues.

Figure 10. Schematic diagram of the experimental design used to evaluate epoxide

assays.

Sample Preparation

37

3.2 General handling issues

Keeping samples stable under experimental conditions was a challenge. Since

lipid oxidation can be initiated by the light and temperature and epoxides degrade

rapidly, all samples are wrapped in aluminum foil and stored frozen under argon until

analyses. Both test samples and epoxide standards were freshly prepared each day.

Only Milli-QTM

purified water (18 M resistivity) was used to prepare reagents.

3.3 HBr (Hydrobromic acid) Assay(AOCS Cd9-57) Methods and Materials

1. Apparatus/Instrumentation/Equipment

A. Argon source - pre-purified compressed Ar, AirGas, Inc. (East Brunswick, NJ)

B. Volumetric burettes

C. Erlenmeyer flasks – 25 ml, 50 ml

D. Analytical balance – M-310 (Denver Instruments, Bohemia, NY)

E. Stir plate and mini stir bar

F. Micro-pipettes – 1000 μl, 200 μl, 10 ul

2. Chemicals/Solvents

A. Hydrogen bromide solution - 33 wt % in acetic acid (Sigma-Aldrich, St. Louis)

B. Acetic acid - ≥ 99.7% ACS reagent, (Sigma-Aldrich, St. Louis, MO)

C. Potassium hydrogen phthalate (KHP) – 99% (MP Biomedical, Solon, OH)

D. 2,3-epoxybutane – >99% (MP Biomedical, Solon, OH)

38

E. 1,2-epoxy-5-hexene – 97% (Sigma-Aldrich, St. Louis, MO)

F. 1,2-epoxy-9-decene – 96% (Sigma-Aldrich, St. Louis, MO)

G. Methyl linoleate – >99% (NU-Chek Prep, Inc., Elysian, MN)

H. Corn oil – supplied by Libra Laboratories, Metuchen, NJ

I. 18 M resistivity pure water –, Milli-Q™ Water Purification System (Millipore,

Billerica, MA)

3. Solutions/Reagents

A. 0.1 N HBr solution: Add 1 mL HBr (33% w/v) to 39 mL glacial acetic acid.

B. KHP primary standard solution (for standardizing the concentration of HBr

solution):

Dry KHP 1-2 hrs at 110 C; cool. Accurately weigh 0.4 0.0001 g KHP and mix it

with 10 mL glacial acetic acid and warm gently to dissolve. Cool to room

temperature.

D. Crystal violet indicator: Dissolve 0.1 g crystal violet in 100 mL glacial acetic acid.

E. Sparge all reagents with argon.

F. Working solutions for epoxides for preparing reaction curves: Add 183 μl

1,2-epoxy-9-decene standard with 10mL acetic acid to make 0.1M solution in test

tub and then dilute stock solutions 1:10 serially to prepare standard solutions ranging

in concentration from 10E-1 to 10E-4 epoxydecene.

39

Concentration C10 1,2-epoxy-9-decene gradient standard solutions

10-1

M 183 ul, C10 1,2-epoxy-9-decene in 10ml acetic acid

10-2

M 1ml 10-1

M + 9ml acetone

10-3

M 1ml 10-2

M + 9ml acetone

10-4

M 1ml 10-3

M + 9ml acetone

G. Add 113 μl 1,2-epoxide-5-hexene in 10mL acetic acid to make 0.1M solution.

and then dilute stock solutions 1:10 serially to prepare standard solutions ranging

in concentration from 10E-1 to 10E-4 epoxyhexene.

Concentration C6 1,2-epoxy-6-dexene gradient standard solutions

10-1

M 113 ul, C6 1,2-epoxy-6-dexene in 10ml acetic acid

10-2

M 1ml 10-1

M + 9ml acetone

10-3

M 1ml 10-2

M + 9ml acetone

10-4

M 1ml 10-3

M + 9ml acetone

H. Add 87μl of 2,3-epoxybutane in 10mL acetic acid to make 0.1M solution and

then dilute stock solutions 1:10 serially to prepare standard solutions ranging

I. in concentration from 10E-1 to 10E-4 epoxybutane.

Concentration C4 2,3-epoxybutane gradient standard solutions

10-1

M 87ul, C4 2,3-epoxybutane in 10ml acetic acid

10-2

M 1ml 10-1

M + 9ml acetone

10-3

M 1ml 10-2

M + 9ml acetone

10-4

M 1ml 10-3

M + 9ml acetone

40

J. Methyl linoleate and corn oil oxidation: 40℃ for 3 days with shaking in an

incubator.

K. Methyl linoleate solutions:

Add 331 ul oxidized methyl linoleate to 10 mL acetic acid (Stock Solution).

Dilute stock solutions 1:10 serially to prepare standard solutions ranging in

concentration from 10E-1 to 10E-4 methyl linoleate.

Concentration methyl linoleate gradient standard solutions

10-1

M 331ul, methyl linoletae in 10ml acetic acid

10-2

M 1ml 10-1

M + 9ml acetone

10-3

M 1ml 10-2

M + 9ml acetone

10-4

M 1ml 10-3

M + 9ml acetone

L. Corn oil solutions:

Add 315 ul oxidized corn oil to 10 mL acetic acid (Stock Solution). Dilute stock

solutions 1:10 serially to prepare standard solutions ranging in concentration

from 10E-1 to 10E-4 corn oil.

Concentration corn oil gradient standard solutions

10-1

M 315ul, corn oil in 10ml acetic acid

10-2

M 1ml 10-1

M + 9ml acetone

10-3

M 1ml 10-2

M + 9ml acetone

10-4

M 1ml 10-3

M + 9ml acetone

41

M. Serial dilution of hydrobromic acid (HBr)

Concentration HBr gradient standard solutions

10-1

M 1ml HBr in 39 ml acetic acid

10-2

M 5ml 10-1

M + 45 ml acetic acid

10-3

M 5ml 10-2

M + 45 ml acetic acid

10-4

M 5ml 10-3

M + 45 ml acetic acid

4. Standardize hydrobromic acid

Titrate the HBr solution with KHP primary solution using no more than 2 drops

crystal violet indicator /50 ml acid. The endpoint is determined as the point at which

a faint purple color persists (all acid has been reacted and pH rises).

Note: The HBr must be standardized daily, and even more frequently when results

become erratic.

5. Determine Oxirane Oxygen %

A. Weigh 0.39~0.41 g diluted epoxide solution into a 25 mL flask with 10 mL

acetic acid (flush with Ar).

B. Put the flask on the stir plate and mini stir bar in the flask.

C. Load the standardized HBr solution in burette (flush with Ar).

D. Use standardized HBr solution to titrate the epoxide solution with stirring

(3drops of crystal violet are added as indicator).

42

E. Titrate until endpoint is reached (when the solution color changes from purple

to blue).

6. Calculations

A. Calculate concentration of HBr solution using the following equation:

B. Determine percentage Oxirane Oxygen:

3.4 NBP (4-p-nitrobenzyl-pyridine) Assay Methods and Materials

1. Apparatus/Instrumentation/Equipment

A. Glass test tubes

B. Micropipets (1000 μl, 200 μl, 10 ul)

C. Spectrophotometer –Varian Cary 50 Bio UV-visible (Varian, Inc, Palo Alto, CA)

D. UV quartz cuvettes – 3.5mL (Sigma-Aldrich, St. Louis, MO)

E. Volumetric flask (10 ml, 25 ml, 100 ml)

F. Plastic squeeze bottle

G. Analytical balance: M-310 (Denver Instruments, Bohemia, NY)

43

2. Chemicals/Solvents

J. 4-p-nitrobenzyl-pyridine – ≥ 98% (Sigma-Aldrich, St. Louis, MO)

K. 2,3-epoxybutane – >99% (MP Biomedical, Solon, OH)

L. 1,2-epoxy-5-hexene – 97% (Sigma-Aldrich, St. Louis, MO)

M. 1,2-epoxy-9-decene – 96% (Sigma-Aldrich, St. Louis, MO)

N. Methyl Linoleate – >99% (NU-Chek Prep, Inc., Elysian, MN)

O. Corn oil – supplied by Libra Laboratories, Metuchen, NJ

P. Potassium hydrogen phthalate (KHP) – 99% (MP Biomedical, Solon, OH)

Q. Potassium carbonate – certified ACS ≥ 98.5% (Fisher Scientific, Pittsburgh, PA)

R. Acetone – Chromasolve®

, for HPLC, ≥99.9% (Sigma-Aldrich, St. Louis, MO)

S. 18 M resistivity pure water –, Milli-Q™ Water Purification System (Millipore,

Billerica, MA)

3. Reagents/solutions

A. 5% NBP reagent: 5 g NBP in 10 mL acetone

B. 0.1 N KHP solution: weigh 2.0422 g KHP in 100 mL Milli-Q water (use hot

water to dissolve KHP)

C. 1 M K2CO3 solution: weigh 8.262 g K2CO3 in 50 mL Milli-Q water

D. 0.1 M 1,2-epoxy-9-decene: pipet 183 μl 1,2-epoxy-9-decene with 10 mL acetone

in a test tube.

44

E. Working solutions for epoxide standard curves: Dilute stock solutions 1:10

serially to prepare standard solutions ranging in concentration from 10E-1 to

10E-6 epoxydecene.

Concentration C10 1,2-epoxy-9-decene gradient standard solutions

10-1

M 183 ul, C10 standard + 10 ml acetone

10-2

M 1 ml 10-1

M + 9 ml acetone

10-3

M 1 ml 10-2

M + 9 ml acetone

10-4

M 1 ml 10-3

M + 9 ml acetone

10-5

M 1 ml 10-4

M + 9 ml acetone

10-6

M 1 ml 10-5

M + 9 ml acetone

F. 0.1 M 1,2-epoxy-5-hexene: pipet 113 μl 1,2-epoxy-5-hexene with 10 mL acetone

in a test tube. Working solutions for epoxides for preparing reaction curves:

Dilute stock solutions 1:10 serially to prepare standard solutions ranging in

concentration from 10E-1 to 10E-6 epoxyhexene.

Concentration C6 1,2-epoxy-5-hexene gradient standard solutions

10-1

M 113ul, C6 standard + 10 ml acetone

10-2

M 1ml 10-1

M + 9 ml acetone

10-3

M 1ml 10-2

M + 9 ml acetone

10-4

M 1ml 10-3

M + 9 ml acetone

10-5

M 1ml 10-4

M + 9 ml acetone

10-6

M 1ml 10-5

M + 9 ml acetone

45

G. 1 M 2,3-epoxybutane: pipet 87 μl 2,3-epoxybutane with 10 mL acetone in a test

tube. Working solutions for epoxides for preparing reaction curves: Dilute stock

solutions 1:10 serially to prepare standard solutions ranging in concentration

from 10E-1 to 10E-6 epoxybutane.

Concentration C4 2,3-epoxybutane gradient standard solutions

10-1

M 87ul, C4 standard + 10 ml acetone

10-2

M 1ml 10-1

M + 9 ml acetone

10-3

M 1ml 10-2

M + 9 ml acetone

10-4

M 1ml 10-3

M + 9 ml acetone

10-5

M 1ml 10-4

M + 9 ml acetone

10-6

M 1ml 10-5

M + 9 ml acetone

H. Methyl linoleate and corn oil oxidation: Both ML and corn oil were oxidized at

40 ℃ in an incubator for 3 days oxidation.

I. ML serial dilutions: Dilute stock solutions 1:10 serially to prepare standard

solutions ranging in concentration from 10E-1 to 10E-6 ML as shown below.

Concentration methyl linoleate gradient standard solutions

10-1

M 331 ul methyl linoleate in 10 ml acetone

10-2

M 1ml 10-1

M ML + 9 ml acetone

10-3

M 1ml 10-2

M ML+ 9 ml acetone

10-4

M 1ml 10-3

M ML + 9 ml acetone

10-5

M 1ml 10-4

M ML + 9 ml acetone

10-6

M 1ml 10-5

M ML + 9 ml acetone

46

J. Corn oil serial dilutions: Add 315 ul corn oil to 10 ml acetone. Dilute stock

solutions 1:10 serially to prepare standard solutions ranging in concentration

from 10E-1 to 10E-6 oil.

Concentration corn oil gradient standard solutions

10-1

M 315ul corn oil in 10ml acetone

10-2

M 1ml 10-1

M + 9ml acetone

10-3

M 1ml 10-2

M + 9ml acetone

10-4

M 1ml 10-3

M + 9ml acetone

10-5

M 1ml 10-4

M + 9ml acetone

10-6

M 1ml 10-5

M + 9ml acetone

4. Epoxide assay

A. Add 1 ml epoxide soln to 1 ml acetone in 15 ml test tube (Blank will contain 2 ml

acetone). (Prepare analyses for each concentration dilution.)

B. Add 1 ml 0.1N KHP to 1 ml 5% NBP reagent in a small test tube or Eppendorf tube.

C. Transfer solution (B) quantitatively to solution (A) for each respective sample and

blank, shake or vortex to mix.

D. Heat test tubes in a boiling water bath for 40 minuts for the reaction to occur

E. Cap tubes with marbles to prevent evaporation.

F. Cool tubes to 40~50 C on ice.

G. Add acetone to bring volume to 9 mL.

47

H. Shaking and return to ice bath to bring solutions to room temperature.

I. Add 1 ml K2CO3, mix well, transfer to optical cuvette, and read the absorbance at

600 nm immediately. The violet color fades over time.

3.5 HPLC-DETC (N,N-Diethyldithiocarbamate) Assay Materials and Methods

1. Appraratus/Instrumentation/Equipment

A. HPLC – LC-10AD (Shimadzu Scientific, Columbia, MD)

B. Autosampler – SIL-10ADvp (Shimadzu Scientific, Columbia, MD)

C. Detector – SPD-M10Avp (Shimadzu Scientific, Columbia, MD)

D. Column – Ultra C18, 5um, 150*4mm (Restek Corp, Bellafonte, PA)

E. HPLC sample bottle – 2 ml

F. Test tubes – 5 ml

G. Micro-pipettes – 1000 μl, 200 μl, 10 ul

H. Analytical balance – M-310 (Denver Instruments, Bohemia, NY)

2. Chemicals/Solvents

A. Methanol – ≥99.8%, Chromasolv® for LC-MS (Sigma-Aldrich, St. Louis, MO)

B. Milli-Q™ Water System, 4 cartridge, 18 M resistivity (Millipore, Billerica, MA)

C. O-Phosphoric acid – 85% HPLC grade (Fisher Chemicals, Pittsburgh, PA)

D. 2,3-epoxybutane – >99% (MP Biomedical, Solon, OH)

48

E. 1,2-epoxy-5-hexene – 97% (Sigma-Aldrich, St. Louis, MO)

F. 1,2-epoxy-9-decene – 96% (Sigma-Aldrich, St. Louis, MO)

G. Methyl Linoleate – >99% (NU-Chek Prep, Inc., Elysian, MN)

H. Corn oil – supplied by Libra Laboratories, Metuchen, NJ

I. Sodium diethyldithiocarbamate trihydrate – ≥99.0%, ACS reagent

(Sigma-Aldrich, St. Louis, MO)

3. Reagents/solutions

A. 0.01 M DETC: 0.0225 g DETC in 10 mL Methanol.

B. 0.02 M 2,3-epoxybutane: add 17.5 μl 1,2-epoxybutane in 10 mL methanol.

C. 0.02 M 1,2-epoxy-5-hexene: add 22.6 μl 1,2-epoxy-5-hexene in 10 mL

methanol.

D. 0.02 M 1,2-epoxy-9-decene: add 36.6 μl 1,2-epoxy-9-decene in 10 mL methanol.

E. DETC and epoxide mixture(10-4

M): 4 mL, 0.01 M DETC + 20 μl, 0.02M

epoxide solution in a test tube.

F. Blank solution: add 4mL, 0.01M DETC in a test tube.

4. Reaction solutions for HPLC analysis

A. Incubate the DETC and epoxide mixture solution: Incubate solution E in 60℃

warm water bath for 20 minutes.

B. Stop the reaction by adding 100 μl phosphoric acid to the mixture.

49

C. Dilute the DETC epoxide mixture for HPLC analysis:

2 mL DETC-epoxide mixture (10-4

M) + 8 mL methanol

D. Serial dilution solution of DETC-epoxide mixture with methanol

Concentration of

DETC-epoxide mixture

for HPLC

20uM DETC-epoxide

mixture added (μl)

Methanol added (μl )

20 μM 1000 0

15 μM 750 250

10 μM 500 500

5 μM 250 750

2.5 μM 125 875

E. Blank solution: 1mL DETC (reagent F)

F. Put 1 mL each DETC-epoxide mixture in vial for HPLC

G. Inject 20 μL of each solution for HPLC analysis

5. Methyl linoleate and corn oil

A. ML and corn oil were oxidized in an incubator under 40℃ for 3 days.

B. Mix 0.112 g DETC and 33 μl methyl linoleate in 10 mL methanol in a test tube

to make 0.01 M solution.

C. Mix 0.112 g DETC and 31 μl corn oil in 10 mL methanol in a test tube to make

0.01 M solution.

D. Incubate test tubes in a warm water incubation 60℃ for 20 minutes.

50

E. Add 200 μl phosphoric acid to the solution to stop the reaction and decompose

excess DETC.

F. Vortex tubes well.

G. Dilute DETC-epoxide mixtures:

Concentration(M) DETC-epoxide added Methanol added

10-2

M - -

10-3

M 1000 μl 10-2

M 9000 μl

7.5 x 10-4

M 6000 μl 10-3

M 2000 μl

5 x 10-4

M 4000 μl 7.5x10-3

M 2000 μl

2.5 x 10-4

M 4000 μl 5x10-3

M 4000 μl

10-4

M 1000 μl 10-3

M 9000 μl

6. Inject 20 l of each sample into the HPLC for separation and analysis.

7. HPLC instrumental conditions

Instrument: Shimadzu

Autosampler: Shimadzu SIL-10AD vp

Injection volume: 20 l

Detector: Shimadzu SPD-M10Avp

Pump: Shimadzu LC-10ADvp*2

Column: Restek Ultra C18 (5 m, 150*4.6 mm)

Gradient:

M03 Mobile Phase: 2.0 mL/min

A: Acetonitrile

B: H2O

51

Time (min) 0 2 10 12 17 19 20 (stop)

ACN (ml/min) 0.80 0.80 1.60 2.00 2.00 0.80 0.80

H2O (ml/min) 1.20 1.20 0.40 0.00 0.00 1.20 1.20

3.6 NMR (Nuclear Magnetic Resonance) Assay Materials and Methods

1. Appraratus/Instrumentation/Equipment

A. Test tubes – 5 ml

B. NMR tubes – standard quartz, 5 mm OD

C. NMR – Varian 400 MHz (Varian, Inc., Palo Alto, CA)

D. Micro-pipettes – 1000 μl, 200 μl, 10 ul

2. Chemicals/Solvents

A. 2,3-epoxybutane – >99% (MP Biomedical, Solon, OH)

B. 1,2-epoxy-5-hexene – 97% (Sigma-Aldrich, St. Louis, MO)

C. 1,2-epoxy-9-decene – 96% (Sigma-Aldrich, St. Louis, MO)

D. Methyl Linoleate – >99% (NU-Chek Prep, Inc., Elysian, MN)

E. Corn oil – supplied by Libra Laboratories, Metuchen, NJ

F. Chloroform-d (CDCl3) – Sigma 99.96%D

3. Reagents/solutions

A. 1 M 1,2-epoxy-9-decene: Add 165 μl epoxide to 0.9 ml CDCl3 in a test tube.

B. Serial dilutions of C10 1,2-epoxy-9-decene from1M to 10-6

M epoxide:

Add C10 1,2-epoxy-9-decene 0.1 ml to 0.9 ml CDCl3 and then dilute stock

52

solutions 1:10 serially to prepare standard solutions ranging in concentration

from 10E-1 to 10E-4.epoxide.

Concentration(M) 1,2-epoxy-9-decene gradient standard solutions

10-1

M 0.1 ml 1 M C10 + 0.9 mL CDCl3

10-2

M 0.1 ml 10-1

M C10 + 0.9 mL CDCl3

10-3

M 0.1 ml 10-2

M C10 + 0.9 mL CDCl3

10-4

M 0.1 ml 10-3

M C10 + 0.9 mL CDCl3

10-5

M 0.1 ml 10-4

M C10 + 0.9 mL CDCl3

10-6

M 0.1 ml 10-5

M C10 + 0.9 mL CDCl3

C. Methyl linoleate (ML) and Corn oil oxidation: ML and Corn oil were oxidized

at 40℃ with shaking in an incubator for 3 days.

D. Serial dilutions of methyl linoleate: Add 5.96 μl oxidized methyl

linoleate to 0.9 ml CDCl3 then dilute stock solutions 1:10 serially to

prepare standard solutions ranging in concentration from 0.02M to

2x10-6

M.lipid.

Concentration methyl linoleate gradient standard solutions

0.02M 5.96μl ML + 0.9ml CDCl3

2x10-3

M 0.1ml 0.02M + 0.9ml CDCl3

2x10-4

M 0.1ml 2x10-3

M + 0.9ml CDCl3

2x10-5

M 0.1ml 2x10-4

M + 0.9ml CDCl3

2x10-6

M 0.1ml 2x10-5

M + 0.9ml CDCl3

53

E. Serial dilutions of corn oil: Add 6.38 μl oxidized corn oil to 0.9 ml CDCl3 then

dilute stock solutions 1:10 serially to prepare standard solutions ranging in

concentration from 0.02M to 2x10-6

M lipid.

Concentration corn oil gradient standard solutions

0.02 M 6.38 μl corn oil + 1ml CDCl3

2x10-3

M 0.1 ml 0.02 M + 0.9 ml CDCl3

2x10-4

M 0.1 ml 2x10-3

M + 0.9 ml CDCl3

2x10-5

M 0.1 ml 2x10-4

M + 0.9 ml CDCl3

2x10-6

M 0.1ml 2x10-5

M + 0.9 ml CDCl3

4. NMR analysis: Add 0.7 mL of each solution above in a NMR tube for 1H NMR

analyses.

5. NMR running condiction

NMR Varian VNMRS 400 MHz

Temperature 25 C

Spin 20 Hz

Solvent Chloroform-d (CDCl3) – Sigma 99.96%D

Times of NMR scan 1H NMR 8 scans

54

4 . RESULTS AND DISCUSSION

4.1 HBr (Hydrobromic acid) Assay

HBr titration provides a means of determining how much epoxy oxygen is present

in a sample, expressed as the percent (%) oxirane oxygen. Although this assay has been

accepted as a standard procedure for several decades, it was the most problematic of all

assays we evaluated. This assay required more than average skill in titration and the end

point (based on color change from purple-violet to blue-green) was difficult to

determine accurately and reproducibly.

Detection range and accuracy. We found the assay was useful only when the epoxide

concentrations were high. The detection range of the HBr assay was 0.0075 M to 0.1 M.

However, the response was not linear below 10 mM epoxide and the response varied

with epoxide structure, increasing with epoxide chain length (Figure 11, Table 4). At

0.1 M concentration, the detected % oxirane oxygen of 1,2-epoxy-9-decene was

slightly above the theoretical value of 0.152, 1,2-epoxy-5-hexene was the most accurate,

and 2,3-epoxybutane was slightly below the theoretical value. At 0.001 M standard, the

experimental result of all 3 epoxide standard tested were about twice the theoretical

value. At 0.0001 M epoxide standard, samples overtitrated by about ten times, probably

due to the 0.1 M HBr prescribed by the standard assay which was designed for high

epoxide concentrations in modified oils. However, at lower concentrations HBr was too

volatile and autoxidized readily so titrations were erratic and inaccurate.

55

C6

C10

Figure 11. Titration curves for reaction of HBr with epoxybutane, epoxyhexene, and

epoxydecene standards.

Table 4. Percent oxirane oxygen detected in different concentrations of epoxydecene,

epoxyhexene and epoxybutane standards by HBr titration.

% of Oxirane Oxygen

Conc (M) Epoxydecene Epoxyhexene Epoxybutane Actual Value

0.1 0.16069 0.15336 0.11335 0.152

0.01 0.01601 0.01514 0.01329 0.0152

0.0075 0.01508 0.01257 0.01135 0.0114

0.005 0.01134 0.00938 0.00832 0.0076

0.0025 0.00687 0.00576 0.00533 0.0038

0.001 0.00326 0.00261 0.00321 0.00152

0.0001 0.00182 0.00146 0.00117 0.000152

C10

C6

C4

56

Application to epoxides in oxidized lipids. To test suitability of the HBr assay

with complex materials such as lipids, the assay was applied to analysis of epoxides in

oxidized methyl linoleate and corn oil. Assessment of this application is hampered by

not having a primary assay to determine actual spoxide concentrations. However, if this

assay is correct, the oils either are barely oxidized or have very low levels (micromolar)

of epoxide products present (Table 5).

Table 5. HBr titration detection of epoxides in oxidized corn oil and methyl linoleate.

% of Oxirane Oxygen

Oil Conc. (M) Corn Oil C10 equiv.(M) Me Linoleate C10 equiv. (M)

0.1 0.0023 0.00033 0.0021 0.0002

0.01 0.0017 0.00009 0.0016 0.00008

0.001 0.0016 0.00008 0.0011 0.00006

0.0001 0.0011 0.00006 0.0025 0.0004

Reproducibility. To test reproducibility of the assay, five replicate analyses of

samples were performed on each of three days and within day and between day

standard deviations (Tables 6-10) and coefficients of variance were calculated (Table

11). Within day reproducibility for the standards was excellent, 3% or lower variability.

Between day variability was slightly higher, but still < 6% for the longer epoxides and

<10% for epoxybutane. The high volatility of epoxybutane contributes to its higher

57

variability. Variability of methyl linoleate was comparable to the standards; corn oil

was slightly higher, reaching 12% between days. Corn oil by its nature is not totally

homogeneous, and it continues to oxidize and degrade even when frozen, so the higher

variability between days is perhaps to be expected.

Table 6. Reproducibility of HBr titration assay for C10 1,2-epoxy-9-decene. Five

replicates per day, three different days.

Within day % Oxirane O

1,2-epoxy-9-decene Day 1 Day 2 Day 3 Between days

0.1M Ave 0.1595 0.1642 0.1591 0.1610

Stdev 0.0016 0.0061 0.0027 0.0028

0.01M Ave 0.0149 0.0182 0.015 0.0160

Stdev 0.0015 0.0018 4E-05 0.0019

0.0075M Ave 0.0150 0.0154 0.0149 0.0151

Stdev 0.0006 0.0002 0.0007 0.0003

0.005M Ave 0.0114 0.0113 0.0113 0.0113

Stdev 0.0004 9E-05 8E-05 3.1E-05

0.0025M Ave 0.0069 0.0069 0.0069 0.0069

Stdev 0.0001 0.0003 0.0002 2.6E-05

0.001M Ave 0.0033 0.0032 0.0033 0.0033

Stdev 0.0002 8E-05 0.0001 6.5E-05

0.0001M Ave 0.0018 0.0018 0.0019 0.0018

Stdev 0.0002 9E-05 0.0002 7.0E-05

58

Table 7. Reproducibility of HBr titration assay for 1,2-epoxy-5-hexene. Five replicates

per day, three different days.

1,2-epoxy-5-hexene Within day % Oxirane O

Conc. (M) Day 1 Day 2 Day 3 Between days

0.1 Ave 0.1534 0.1536 0.1531 0.1534

Stdev 0.0031 0.002 0.0035 0.0003

0.01 Ave 0.0128 0.0156 0.0171 0.0151

Stdev 0.0013 0.0017 0.0002 0.0022

0.0075 Ave 0.0130 0.0123 0.0123 0.0126

Stdev 0.0005 0.0007 0.0007 0.0004

0.005 Ave 0.0094 0.0093 0.0095 0.0094

Stdev 0.0001 0.0001 0.0002 0.0001

0.0025 Ave 0.0058 0.0058 0.0057 0.0058

Stdev 0.0001 0.0001 2.00E-05 0.0000

0.001 Ave 0.0023 0.0034 0.0022 0.0026

Stdev 4.00E-05 3.00E-05 1.00E-05 0.0007

0.0001 Ave 0.0015 0.0015 0.0014 0.0015

Stdev 1.00E-05 2.00E-05 2.00E-05 0.0000

59

Table 8. Reproducibility of HBr titration assay for C4 2,3-epoxbutane.

Five replicates per day, three different days.

2,3-Epoxybutane Within day % oxirane O

Conc. (M) Day 1 Day 2 Day 3 Between days

0.1 Ave 0.1217 0.0943 0.1241 0.11335

Stdev 0.0086 0.0481 0.0016 0.01654

0.01 Ave 0.0131 0.0134 0.0134 0.01329

Stdev 0.0009 0.0002 6.00E-05 0.00016

0.0075 Ave 0.0112 0.0116 0.0113 0.0135

Stdev 0.0002 2.00E-05 2.00E-05 0.00018

0.005 Ave 0.0083 0.0084 0.0083 0.00832

Stdev 0.0001 0.0002 0.0002 4.03E-05

0.0025 Ave 0.0054 0.0053 0.0053 0.00533

Stdev 0.0001 8.00E-05 3.00E-05 2.63E-05

0.001 Ave 0.0047 0.0027 0.0022 0.00321

Stdev 0.0008 3.00E-05 2.00E-05 0.00135

0.0001 Ave 0.0011 0.0012 0.0012 0.00117

Stdev 4.00E-05 0.0001 9.00E-06 4.23E-05

60

Table 9. Reproducibility of HBr titration assay for oxidized methyl linoleate.

Five replicates per day, three different days.

ML Within day % Oxirane O

Conc. (M) Day 1 Day 2 Day 3 Between days

0.1 Ave 0.0018 0.0024 0.0021 0.0021

Stdev 9.00E-06 9.00E-06 9.00E-05 0.0003

0.01 Ave 0.0018 0.0014 0.0016 0.0016

Stdev 2.00E-05 0.0052 0.0001 0.0002

0.001 Ave 0.0011 0.0011 0.0011 0.0011

Stdev 7.00E-06 4.00E-05 5.00E-05 0.0000

0.0001 Ave 0.0024 0.0026 0.0024 0.0025

Stdev 1.00E-04 0.0002 0.0001 0.0001

Table 10. Reproducibility of HBr titration assay for oxidized corn oil. Five replicates

per day, three different days.

Corn Oil Within day % Oxirane O

Conc. (M) Day 1 Day 2 Day 3 Between days

0.1 Ave 0.0021 0.0023 0.0024 0.0023

Stdev 5.00E-05 0.0002 4.00E-05 0.0001

0.01 Ave 0.0017 0.0017 0.0018 0.0017

Stdev 0.0001 1.00E-05 3.00E-05 0.0001

0.001 Ave 0.0013 0.0013 0.0022 0.0016

Stdev 4.00E-06 5.00E-05 0.0018 0.0005

0.0001 Ave 0.001 0.0011 0.0012 0.0011

Stdev 7.00E-06 9.00E-05 5.00E-05 0.0001

61

Table 11. Coefficients of variation for HBr titration of standard epoxides.

Coefficient of Variation (%)

Within day Between Days

C4 3.38 9.17

C6 1.87 6.37

C10 2.13 3.10

Corn oil 5.64 12.93

Me Linoleate 3.89 7.93

Handling and Precautions. A number of factors contribute to increasing

inaccuracy of the HBR assay in detecting lower concentrations of epoxides:

(1) Innate instability and reactivity of hydrobromic acid (HBr). Typically,

acid with low pKa values dissociate more readily and thus donate H+ to the

surroundings. Acids with pKa values lower than -2 are considered strong acids.

Hydrobromic acid has a pKa = -9 which means it is a very strong acid and is nearly

totally dissociated under most conditions. The free bromine and H+ ions then evaporate

readily from the very beginning of the titration, contributing to variability and to

apparent over-titrations because not all the hydrobromic acid leaving the burette

reaches the epoxide. Volatility increases as the chain length decreases. Lower volatility

thus may partially contribute to the apparent increased reactivity with HBr as chain

length of the epoxide increases.

Also, HBr oxidizes readily, so less actual HBr is present in the titrant delivered to

62

the epoxide. As a consequence, oxirane values determined in air are consistently lower

than when the solutions and headspace are sparged with argon (data not shown). To

reduce variability and improve accuracy, we thus routinely use argon to flush the

reagents, solutions, and flasks headspace before and during titration.

Finally, HBr reaction is not specific for epoxides; it also reacts with other

substances. This is not a problem when pure compounds are being analyzed, but may be

a significant limitation when titrating oxidized lipids that potentially contain a wide

variety of oxidation products. HBr reacts with β –unsaturated ketones, cyclopropenes,

conjugated dienols, soaps, and other oxygenated secondary products of lipid oxidation.

Alternate targets compete with epoxides for HBr and may react in addition to epoxides,

so whether the net effect is aynergistic or antagonistic is not clear. At any rate, results in

complex systems (as opposed to pure materials) are questionable.

2) Poor definition in endpoint. The color does not change completely and

instantly at the endpoint of HBR titration of epoxides. The indicator, crystal violet, has

a purple violet color. As hydrobromic acid adds to the epoxide during titration, the color

gradually fades and turns into blue. It is not easy to visually determine the exact point at

which the color changes. In addition, the perception of color change is subjective and

varies tremendously among individuals.

63

4.2 NBP (4-p-nitrobenzyl-pyridine) Assay

Procedures in original references for this method (Hammock et al. 1974, Agarwal

et al. 1979) were mostly qualitative so required some adaptation to make them

quantitative. The first problem encountered was determining the optimum heating time

for the reaction. A range of heating times was tested, from 20 to 50 minutes (Figure 12).

40 minutes provided the most consistent and linear results so was selected as the

standard time for all tests reported here.

Effect of epoxide structure on reaction response. The second problem was

quantitation. Clearly, a standard curve constructed from the epoxide of interest can

provide a basis for quantitation. However, as shown in Figures 12 and 13, the NBP

reaction response varies markedly with epoxide structure, so an authentic epoxide

standard is required for every epoxide analyzed. For lipids, such standards are very

difficult to obtain commercially, and multiple epoxides are generated during lipid

oxidation and they change with reaction conditions, so even determining which

epoxides should be synthesized as standards is complicated. Thus, the assay can report

relative concentrations and demonstrate changes over time, but it cannot provide the

absolute quantitation of epoxides that is necessary for mass balance and comparison of

alternate reactions in lipid oxidation. One option is to use epoxydecene or any longer

epoxide that is commercially available as a standard and then report epoxide

64

Figure 12. Effects of heating time on NBP reaction response of three standard epoxides.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0 0.2 0.4 0.6 0.8 1

2,3-Epoxybutane

20 mins

30 mins

40 mins

50 mins

0.0

0.5

1.0

1.5

2.0

2.5

3.0

0 0.02 0.04 0.06 0.08 0.1

1,2-Epoxy-5-hexene

0

1

2

3

4

0 0.003 0.006 0.009 0.012

1,2-Epoxy-9-decene

Epoxide conc. (M)

Ab

so

rba

nce

6

00

nm

65

Figure 13. NBP reaction response curves for epoxybutane, epoxyhexene, and

epoxydecene standards. (40 minutes standard reaction time).

concentrations as equivalents of the standard used.

Figures 12 and 13 show clearly that epoxides with different chain lengths (or R-

groups), NBP reaction response can vary by orders of magnitude. 1,2-epoxy-9-decene

had the greatest response by far, reaching the detection limit of the spectrophotometer

at a concentration of 0.1M. Epoxyhexene also exceeded useful detection limits at 0.1 M

but the absorbance levels were lower. In contrast, epoxybutane did not reach optical

limts even at 1 M concentrations.

The strong effects of epoxide R group on reaction with NBP can create significant

problems in complex materials with unknown and potentially multiple epoxide

y = -1456.5x2 + 239.66x + 0.314 R² = 0.9711

y = 71.021x2 + 38.556x + 0.0958 R² = 0.9874

y = 0.3849x + 0.0769 R² = 0.9968

0

2

4

6

8

10

12

0 0.2 0.4 0.6 0.8 1

Ab

so

rban

ce

(6

00

nm

)

Concentration (M)

Epoxyhexene

Epoxybutane

Epoxydecene

66

structures. Some of this difference may be attributed to high volatility in the shorter

chain epoxides. Epoxybutane is particularly volatile and thus difficult to handle and

prepare samples reproducibly. Another explanation for the more rapid and complete

reaction of long chain epoxides with NBP may arise from inductive effects of the side

chains on charge distribution in the epoxy ring. During the reactions, the acyl chains (R

groups) next to epoxy ring can act as electron donating groups that increase electron

density on the oxirane ring. This creates a polar gradient in the epoxy ring, with a partial

negative charge on the oxirane oxygen and a partial positive charge on one oxirane

carbon. Since larger molecule epoxide C10 1,2-epoxy-9-decene has a longer acyl chain

which will push orbital electrons much harder to the oxirane oxygen result in a more

partial negative charge on oxirane oxygen and a more partial positive charge on oxirane

carbon. At the same time, the NBP reagent possesses an extra lone pair of electrons

which are attracted by the positive charge of oxirane carbon, as shown below:

Because longer chains are more strongly electron withdrawing, the oxirane carbon is

more strongly positive, and NBP thus reacts with these more readily. Hence, reactivity

increases in the order C4 < C6 < C10 epoxides.

Detection limits for assay. Useful detection ranges for the three standard

67

epoxides are presented in Table 12. Epoxybutane can be detected over the range 10 M

to 1 M, but the response is thrd order polynomial, not linear, indicatint that

concentration is not the only factor affecting reactivity. C6 and C10 epoxides can both

be detected as low as 1 M, but different upper limits – 100 mM for C6 and 50 mM for

C10 where the absorbance levels generated exceed the spectrophotometer limits.

The Cary 50 spectrophotometer used in this study uses a xenon lamp and has a quoted

optical limit of 10, although results above 3 are increasingly noisy. However, for most

spectrophotometers with monochomators and deuterium and hydrogen lamps, the

optical limit for accurate reading is an absorbance of 1. Thus, for most laboratories the

upper detection limits for the NBP assay are substantially lower than those reported

here.

Table 12. Active detection ranges for NBP assay.

Low detection limt High detection limit

C4 10 uM 1 M

C6 1 uM 0.1 M

C10 1 uM 0.05 M

ML 1 uM (oil) 1 M (oil)

Corn oil 1 uM (oil) 1 M (oil)

Linearity of reaction response. Advantages of increased reactivity at longer

chain lengths are counterbalanced by decreasing linearity of response. The reaction

68

response was low. Below 1 M there was no detectable reaction. Response was linear

over the range of 1 M to 1 M for epoxybutane, but became increasingly non-linear

(polynomial) as the R chain length increased. Responses for epoxyhexene were linear

only at low concentrations, between 10-5

M and 10-6

M, but epoxydecene response

remained polynomial even at the lowest concentrations (Figure 14). This suggests that

there other factors such as steric accessibility or competing side reactions controlling

the reactivity, in addition to inductive effects. In the same low concentration range,

epoxybutane response was increasingly variable, reflecting the handling and volatility

problems already noted.

When test concentrations span several orders of magnitude (0.1M to 10-6

M),

semilog plots are sometimes more revealing about reaction patterns than standard

concentration plots. Semilog plots of the low epoxide concentration range (10-5

M to

10-6

M) showed a linear relationship between log [epoxide] and absorbance (reaction

response) for epoxydecene, but third order and second order quadratic relationships for

epoxyhexene and epoxybutane, respectively (Figure 15). Hence, reactions of the three

epoxides are controlled by different factors. Epoxydecene has a power (log)

relationship with concentration, epoxydecene has a direct concentration relationship,

and epoxybutane has other factors controlling reaction at low concentrations.

69

y = 2E-08x3 - 2E-06x2 + 0.0001x + 0.0677 R² = 0.9977

0.066

0.068

0.070

0.072

0.074

0.076

0.078

0.080

0 20 40 60 80 100

Ab

so

rban

ce (

600 n

m)

Epoxide Concentration (M)

Figure 14. NBP reaction response at very low concentrations of epoxyhexene and

epoxydecene standards (top) and epoxybutane (bottom).

y = -3E-05x2 + 0.0006x + 0.0693 R² = 0.9737

y = 0.0003x + 0.0693 R² = 0.9919

0.068

0.069

0.070

0.071

0.072

0.073

0.074

0.075

0 2 4 6 8 10

Ab

so

rban

ce (

600 n

m) epoxydecene

epoxyhexene

70

Figure 15. Log epoxide concentration vs absorbance (reaction response) for NBP assay.

Top: epoxydecene and epoxyhexene. Bottom: epoxybutane.

y = 0.0028x + 0.0698 R² = 0.994

y = 0.0063x3 - 0.0066x2 + 0.0032x + 0.0696 R² = 0.9999

0.068

0.069

0.070

0.071

0.072

0.073

0.074

0.075

0 0.2 0.4 0.6 0.8 1

Ab

so

rban

ce (

600 n

m)

Log epoxide concentration (M)

epoxydecene

epoxyhexene

y = 0.0116x2 - 0.0197x + 0.0785 R² = 0.9926

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0.18

0.20

0 1 2 3 4 5

Ab

so

rban

ce (

600 n

m)

Log epoxide concentration (M)

epoxybutane

71

Application to epoxides in oxidized lipids. Similar results were found in methyl

linoleate and corn oil oxidized at 40 C for 3 days (Figure 16). The concentration range

between 1M and 10-6

M oil was examined. Very little epoxide was detected in methyl

linoleate, but substantially more was detected in corn oil. The absorbance in corn oil

suggests a concentration of about 1 mM epoxydecene equivalents, while methyl

linoleate is less than 0.5 mM epoxydecene equivalents. These values are a bit higher

than detected by the HBR assay (0.33 and 0.20 mM for corn oil and ML, respectively)

consistent with the greater sensitivity of this assay. As with the HBr assay, reactions

were biphasic with epoxide concentration, possibly reflecting steric accessibility

limitations at higher epoxide/oil concentrations.

To examine the concentration relationships more closely, semilog plots of the

reaction responses were plotted (Figure 17). Rather than simplifying the relationships

to linear, these plots increased the complexity to third order. Confirming that factors

other than oil concentration are contributing to the reaction response. These factors may

be physical, such as increased viscosity slowing diffusion of the NBP to lipid sites,

steric hindrance by the long acy chains or triacylglycerols structure, or alterations in

NBP solubility. Or the factors may be chemical, related to the structure and functional

groups present in the lipids. More research will be needed to elucidate the reasons for

this variable reactivity..

72

Figure 16. NBP response curves for oxidized methyl linoleate and corn oil. Top: Full

concentration range tested. Bottom: lower concentration range.

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0 0.2 0.4 0.6 0.8 1

Ab

so

rban

ce (

600 n

m)

Oil Concentration (M)

Corn oil

Me Linoleate

y = 0.0667x0.0309 R² = 0.9578

y = 0.0657x0.0741 R² = 0.9301

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0 200 400 600 800 1000

Ab

so

rban

ce (

600 n

m)

Concentration (uM)

Me Linoleate

Corn oil

73

Figure 17. Log ML and corn oil concentration vs absorbance (reaction response) for

NBP assay. Top: full concentration range tested. Bottom: Lower concentration range

(<6 M).

Reproducibility of assay. Five replicates of all samples were analyzed on each of

y = 0.0008x2 + 0.0022x + 0.068 R² = 0.9873

y = -0.0024x3 + 0.0153x2 - 0.0096x + 0.0691 R² = 0.9984

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0 1 2 3 4 5 6

Ab

so

rban

ce (

600n

m)

Log Oil Concentration (M)

y = 9E-05x3 - 6E-05x2 + 0.004x + 0.0674 R² = 0.9886

y = 0.0038x3 - 0.0189x2 + 0.034x + 0.065 R² = 0.9917

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

0 1 2 3 4 5 6

Ab

so

rban

ce (

600n

m)

Log Oil Concentration (M)

Corn oil

Me linoleate

74

three days to test reproducibility of the assay. Coefficients of variation for the assay are

presented in Table 13. Results for epoxide standards are presented in Tables 14-16;

results for methyl linoleate and corn oil are in Tables 17 and 18. Reproducibility of the

assay with standard epoxides was exceptionally good -- 1-2% within days and 2-3%

between days. Variability with methyl linoleate was only slightly higher, 3-4% within

and between days, respectively. Variability of the assay with corn oil was significantly

higher – 7% within day and almost 12% between days. This probably reflected the

innate inhomogeneity of the material, or possibly sampling position, rather than

progressive oxidation since the epoxides di not increase each successive day. This is

still good reproducibility, but with some work on standardizing sampling methods the

variability could be reduced to less than 5% within days and 10% between days.

Table 13. Coefficients of variation for the NBP detection of epoxides in standards and

oxidized lipids. Averaged over all concentrations and days

average COV(%)

Within day Between days

C10 1.53 2.73

C6 1.12 1.87

C4 1.23 2.67

Corn oil 7.32 11.84

ML 3.21 4.17

75

Table 24. Reproducibility of NBP assay for 1,2-epoxy-9-decene standard. Five

replicates per day, three different days.

1,2-epoxy-9-decene Within day (A 600 nm)

Conc. (M) Day1 Day2 Day3 Between Days

0.1 Ave 10 10 10 10

Stdev 0 0 0 0

0.075 Ave 10 10 10 10

Stdev 0 0 0 0

0.05 Ave 7.875 7.7145 6.657 7.4154

Stdev 2.91 3.1365 3.055 0.6619

0.025 Ave 6.443 6.3835 6.384 6.4035

Stdev 3.248 3.3014 3.301 0.0341

0.01 Ave 4.046 4.2131 4.095 4.1179

Stdev 0.058 0.0829 0.054 0.0860

0.001 Ave 0.483 0.5103 0.497 0.4967

Stdev 0.005 0.0114 0.002 0.0137

0.0001 Ave 0.102 0.1087 0.096 0.1021

Stdev 0.008 0.0007 0.004 0.0066

0.00001 Ave 0.072 0.0734 0.073 0.0727

Stdev 4.00E-04 0.0011 6.00E-04 0.0008

7.5E-06 Ave 0.072 0.0733 0.071 0.0721

Stdev 7.00E-04 0.0007 5.00E-05 0.0011

5E-06 Ave 0.071 0.0722 0.072 0.0717

Stdev 6.00E-04 0.0008 4.00E-04 0.0005

2.5E-06 Ave 0.071 0.0718 0.07 0.0709

Stdev 7.00E-04 0.0003 0.001 0.0008

1E-06 Ave 0.068 0.0714 0.069 0.0698

Stdev 0.003 0.0002 2.00E-04 0.0015

76

Table 15. Reproducibility of NBP assay for 1,2-epoxy-5-hexene standard. Five

replicates per day, three different days.

1,2-epoxy-5-hexene Within day A 600 nm

Conc. (M) Day 1 Day 2 Day 3 Between Days

0.1 Ave 4.643 4.6834 4.502 4.6092

Stdev 0.072 0.1792 0.003 0.0955

0.075 Ave 3.661 3.5916 3.663 3.6385

Stdev 0.087 0.0352 0.008 0.0407

0.05 Ave 1.801 1.7546 1.763 1.7729

Stdev 0.015 0.0002 0.004 0.0248

0.025 Ave 1.371 1.375 1.372 1.3729

Stdev 0.006 0.0004 4.00E-04 0.0019

0.01 Ave 0.594 0.604 0.588 0.5953

Stdev 0.003 0.013 3.00E-04 0.0081

0.001 Ave 0.13 0.1156 0.12 0.1219

Stdev 6.00E-04 0.002 3.00E-04 0.0076

0.0001 Ave 0.096 0.0909 0.085 0.0907

Stdev 5.00E-04 0.0019 6.00E-04 0.0055

0.00001 Ave 0.072 0.0729 0.072 0.0725

Stdev 2.00E-04 0.0008 3.00E-04 0.0004

7.5E-06 Ave 0.071 0.072 0.072 0.0716

Stdev 7.00E-04 0.0009 5.00E-04 0.0004

5E-06 Ave 0.069 0.0716 0.071 0.0708

Stdev 8.00E-05 0.0003 4.00E-04 0.0013

2.5E-06 Ave 0.069 0.071 0.07 0.0702

Stdev 2.00E-04 0.0002 7.00E-04 0.0008

1E-06 Ave 0.069 0.07 0.07 0.0696

Stdev 1.00E-04 0.0002 4.00E-04 0.0005

77

Table 16. Reproducibility of NBP assay for 2,3-epoxybutane standard. Five replicates

per day, three different days.

2,3-epoxybutane Within day A 600 nm

Conc. (M) Day1 Day2 Day3 Between Days

0.1 Ave 0.448 0.4502 0.466 0.455

Stdev 0.018 0.0261 0.005 0.0098

0.075 Ave 0.372 0.3767 0.368 0.3721

Stdev 5.00E-04 0.0003 0.003 0.0045

0.05 Ave 0.263 0.2729 0.262 0.266

Stdev 0.003 0.0014 4.00E-04 0.0060

0.025 Ave 0.193 0.1951 0.17 0.186

Stdev 0.002 0.0015 0.046 0.0143

0.01 Ave 0.112 0.1223 0.121 0.1183

Stdev 8.00E-04 0.0002 0.001 0.0056

0.001 Ave 0.089 0.0927 0.094 0.0919

Stdev 0.002 0.0008 4.00E-04 0.0030

0.0001 Ave 0.077 0.0828 0.087 0.0822

Stdev 0.001 0.0003 7.00E-04 0.0048

0.00001 Ave 0.075 0.0774 0.079 0.0771

Stdev 3.00E-04 0.0006 4.00E-04 0.0028

7.5E-06 Ave 0.071 0.072 0.072 0.0718

Stdev 6.00E-04 0.0003 6.00E-04 0.0004

5E-06 Ave 0.069 0.0698 0.07 0.0697

Stdev 1.00E-04 0.0009 5.00E-04 0.0002

2.5E-06 Ave 0.069 0.0693 0.07 0.0695

Stdev 1.00E-04 0.0004 6.00E-04 0.0003

1E-06 Ave 0.069 0.07 0.069 0.0693

Stdev 2.00E-04 0.0011 4.00E-04 0.0007

78

Table 17. Reproducibility of NBP assay for oxidized methyl linoleate. Five replicates

per day, three different days.

Methyl linoleate Within day A 600 nm

Conc. (M) Day1 Day2 Day3 Between Days

1 Ave 0.097 0.1103 0.119 0.1088

Stdev 0.003 0.0008 0.0007 0.0112

0.1 Ave 0.094 0.0988 0.102 0.0981

Stdev 0.0001 0.0005 0.0004 0.0038

0.01 Ave 0.087 0.0825 0.088 0.0858

Stdev 0.002 0.00008 0.0004 0.0030

0.001 Ave 0.085 0.0811 0.086 0.084

Stdev 0.00007 0.0011 0.0003 0.0025

0.0001 Ave 0.072 0.0757 0.079 0.0757

Stdev 0.0003 0.00005 0.0004 0.0037

0.00001 Ave 0.07 0.0699 0.072 0.0704

Stdev 0.0007 0.0004 0.0004 0.0012

1E-06 Ave 0.067 0.0674 0.069 0.0679

Stdev 0.0006 0.0005 0.0001 0.0014

Table 18. Reproducibility of NBP assay for oxidized corn oil. Five replicates per day,

three different days.

Corn Oil Within day

Conc. (M) Day1 Day2 Day3 Between Days

1 Ave 0.344 0.4765 0.378 0.3995

Stdev 0.003 0.0146 1.00E-04 0.0689

0.1 Ave 0.238 0.2323 0.244 0.2381

Stdev 0.032 0.0144 0.002 0.0061

0.01 Ave 0.133 0.0933 0.139 0.1217

Stdev 0.008 0.0002 3.00E-04 0.0248

0.001 Ave 0.126 0.0898 0.121 0.1121

Stdev 0.002 0.0008 8.00E-04 0.0195

0.0001 Ave 0.113 0.0745 0.092 0.0932

Stdev 0.008 0.0005 2.00E-04 0.0192

0.00001 Ave 0.07 0.0714 0.073 0.0715

Stdev 6.00E-04 0.0004 2.00E-04 0.0017

1E-06 Ave 0.069 0.0712 0.068 0.0694

Stdev 0.001 0.0003 0.002 0.0017

79

Handling and Precautions. The following factors strongly affect experimental

results with the NBP assay:

1. Freshness of the reagent. The NBP reagent must be freshly made to provide

accurate experimental results.

2. Handling issues. First is the degree of vortexing in the final step when potassium

carbonate (K2CO3) is added to the epoxide standard dissolved in the acetone. The

purple or violet color begins to form immediately after the potassium carbonate is

added to the standard. However, the epoxide-NBP-carbonate mixture must be

vortexed before the UV absorbance is read to avoid a color gradient within the tube

( light violet color on top and darker violet color in the bottom of the tube). This

arises because the potassium carbonate is dissolved in the water and epoxide is

dissolved in the non-aqueous acetone. Thus, the samples must be vortexed

vigorously to evenly distribute the violet color.

80

4.3 HPLC-DETC (N,N-diethyldithiocarbamate) Assay

Combination of HPLC to analyze DETC-epoxide complexes provided sensitive

detection of individual epoxides and also completely separated mixtures of epoxides

(Figure 18). C6 and C10 epoxides gave comparable responses, while epoxybutane

response was considerably lower. This method offers the additional advantage of

detecting trace epoxides not expected in the sample (note minor peaks int Figure 18).

Aswill be shown in subsequent discussion, the method is less subject to variation in

epoxide structure, has excellent linearity and acceptable reproducibility, so is suitable

for quantitation of mixed epoxides of unknown structures.

Figure 18. HPLC chromatogram of a 1:1:1 mixture of epoxy butane, hexane, and

decene (top). Bottom: Reaction response is linear for micromolar concentrations.

epoxydecene

epoxyhexene

epoxybutane

“X” “X”

y = 6281.6x + 1482.9 R² = 0.9899

0

20000

40000

60000

80000

100000

120000

0 5 10 15 20

Peak A

rea

Epoxide conc (M)

81

Detection limits. The HPLC-DETC assay was conducted over the concentration

range from 2.5 μM to 0.01 M. This assay requires excess DETC to insure full

interaction with the epoxides. Thus, analyzing higher concentrations of epoxide

requires that additional DETC be added to the reaction mixture. However, high

concentrations of DETC precipitate in the HPLC column and can crash the entire

instrument. For this reason, the practical upper detection limit of the assay is around

0.01 M epoxide. For the short chain epoxide, epoxybutane, the lowest detection limit

was 20 μM – about ten times higher than epoxyhexene and epoxydecene (2.5 μM). This

lower sensitivity for epoxybutane was consistent with observations in the two assays

previously discussed, so all three assays show some similar behaviors. As with the NBP

and HBr assays the reasons for lower reaction of epoxybutane are probably loss of some

of the reactants due to volatility, as well as lower inductive effect from fewer methylene

groups on the carbon back bone. The resulting decreased polarity on the epoxy ring

reduces reaction with DETC relative to mid-chain epoxyhexene and long chain

epoxydecene.

Effects of epoxide structure. Interestingly, in this assay, although epoxybutane

still showed much lower reaction, epoxy hexene and decene exhibited comparable

reactivity in the higher concentration ranges, and epoxyhexene was only slightly less

reactive in the lower concentration ranges (Figure 19). Thus, the DETC reaction is

82

Figure 19. HPLC-DETC reaction response curves (reported as peak areas) for

epoxybutane, epoxyhexene, and epoxydecene standards. Top: full concentration range.

Circle denotes inflection region. Bottom: lower concentration range.

0

5000000

10000000

15000000

20000000

25000000

30000000

35000000

0 2000 4000 6000 8000 10000

Are

a u

nd

er

peak (

mA

u)

Epoxide concentration (uM)

epoxybutane

epoxydecene

epoxyhexene

y = 224.35x - 267.28 R² = 0.9989

y = 5666.1x - 1479.5 R² = 0.9991

y = 6214.5x - 329.91 R² = 0.998

0

20000

40000

60000

80000

100000

120000

140000

0 20 40 60 80 100

are

a u

nd

er

pe

ak

(m

Au

)

Epoxide concentration (uM)

epoxybutane

epoxyhexene

epoxydecene

83

more independent of epoxide structure than the HBr and NBP assays. This is a

significant advantage when analyzing complex systems.

Linearity of response. The overall reponse curve is biphasic, with an inflection

point at or below 1 mM. Above and below this point the curves are remarkably linear

(Figure 19), and in this regard the DETC assay is superior to the HBR and NBP assays

for quantitation. As noted above, slopes are significantly steeper in the low

concentration range, then decrease at higher concentrations.

Application to oxidizing lipids. Corn oil and methyl linoleate incubated for 3

days at 40 C were reacted with DETC and applied to the HPLC column. Contrary to

what was observed with the HBr and NBP assays, this assay found substantially higher

epoxide concentrations in the methyl linoleate than the corn oil (Figure 20). The four

epoxide assays were run back to back on the oils to keep the oils the same for each assay.

While some small changes may be expected between assays, such a massive reversal is

unlikely. Thus, the reversal must arise from differences in the way the methyl linoleate

and triacylglycerols react with the DETC. One possibility is that epoxides on the

triacylglycerols have hindered access to the DETC and hence show lower reaction than

the free ester.

The oils showed reaction patterns comparable to the epoxide standards, with

biphasic curves and inflections at or below 1 mM (Figure 20). Expansion of the lower

84

Figure 20. HPLC-DETC response curves for oxidized methyl linoleate and corn oil.

Top: full concentration range. The circle denotes the inflection region between the two

phases. Bottom: lower concentration range.

y = 125256ln(x) - 618524 R² = 0.9639

y = -0.003x2 + 77.562x + 1563 R² = 0.999

0

100000

200000

300000

400000

500000

600000

0 2000 4000 6000 8000 10000

are

a u

nd

er

peak(m

Au

)

Oil Concentration (uM)

Me linoleate

Corn oil

Methyl linoleate

Corn oil

y = 0.0795x2 + 204.99x - 10387 R² = 0.9889

y = -0.0403x2 + 118.2x - 5484.4 R² = 0.9623

0

50000

100000

150000

200000

250000

300000

0 200 400 600 800 1000

are

a u

nd

er

pe

ak

(mA

u)

Concentration (uM)

85

concentration region and calculation of the best fit curves shows that the relationship of

oil concentration is more quadratic (second order polynomial) than linear. This

indicates that factors other than concentration affect the reaction. However, the R2

values for linear regression curves are high enough (0.9842 and 0.9445 for oxidized

methyl linoleate and corn oil, respectively) that the response curves can be used for

quantitation with minimal error.

Reproducibility of epoxide detection by the assay. Coefficients of variation

are presented in Table 19; within and between day averages and standard deviations for

each standard and the oils are presented in Tables 20-24. This assay had good

reproducibility with epoxide standards, well within acceptable limits for quantitative

analyses, but the variability of results was a bit higher than the NBP and HBr assays –

3-5% within day and 5-7% between days. However, performance with methyl linoleate

and corn oil had some problems. Variation within day was 11-18% and between days

was 16-28%, with the higher values arising from the corn oil. This apparent variability

is most likely attributable to the greated sensitivity, detecting M rather than mM

epoxides. Hence, tiny changes make a large difference in epoxide detected. A second

source of variability may be introduced by the HPLC separations. This would apply

also to the standards and account for the higher variability there.

86

Table 19. Coefficients of variation for HPLC-DETC assay of standard epoxides and

oxidized lipids, averaged over all concentrations and days.

average COV(% )

Within day Between days

C10 3.32 5.86

C6 5.52 7.29

C4 3.28 5.22

Corn oil 17.62 28.03

ML 10.66 16.07

87

Table 20. Reproducibility of HPLC-DETC assay for epoxybutane. Five replicates per

day, three different days.

2,3-epoxybutane Within Day (Peak area)

Conc (uM) Day1 Day2 Day3 Between Days

10000 Ave 10290486 10293769 10300000 10294752

Stdev 31211 10235 10613 4833

1000 Ave 1118251 1116903 1116070 1117074

Stdev 1162 1118 2264 1100

100 Ave 19818 21868 24833 22173

Stdev 2660 3384 2394 2521

75 Ave 15872 16250 18133 16752

Stdev 2224 1766 491 1211

50 Ave 10294 10347 11110 10584

Stdev 1283 392 665 456

20 Ave 4217 4142 4807 4389

Stdev 343 265 296 364

88

Table 21. Reproducibility of HPLC-DETC assay for C6 1,2-epoxy5-hexene.

Five replicates per day, three different days.

1,2-epoxy-5-hexene Within Day (Peak area)

Concentration (uM) Day1 Day2 Day3 Average

10000 Ave 31161695 31141590 31197363 31166883

Stdev 63879 62491 76735 28246

1000 Ave 7346637 7388211 7386759 7373869

Stdev 56715 168439 82423 23595.01662

100 Ave 746736 725029 745802 739188.8

Stdev 30953 2411 8181 12271

20 Ave 104278 108404 118875 110519

Stdev 7330 6158 8932 7524

15 Ave 81539 81140 93554 85410.81

Stdev 2804 2270 8251 7054

10 Ave 52371 52514 60360 55081.92

Stdev 3738 3726 2050 4571

5 Ave 25125 23293 32344 26920.63

Stdev 1934 1420 2182 4785

2.5 Ave 11455 10739 14228 12140.53

Stdev 946 1118 873 1842

89

Table 22. Reproducibility of HPLC-DETC assay for C10 1,2-epoxy-9-decene.

Five replicates per day, three different days.

1,2-epoxy-9-decene Within Day

Concentration (uM) Day1 Day2 Day3 Between Day

10000 Ave 30541073 30595281 30656659 30597671

Stdev 246178 192516 11585870 57830

1000 Ave 7519075 7534257 7627924 7560419

Stdev 103969 160221 2883175 58952

100 Ave 686551 703687 822188 737475

Stdev 24923 18055 310759 73862

20 Ave 132169 117481 115785 121812

Stdev 29999 8893 44796 9010

15 Ave 95061 94678 97900 95879

Stdev 11381 6053 37197 1760

10 Ave 68412 56938 59426 61592

Stdev 17415 3413 22500 6036

5 Ave 35202 28922 30312 31479

Stdev 15123 826 11532 3298

2.5 Ave 13455 14643 13444 13847

Stdev 998 962 5243 689

90

Table 23. Reproducibility of HPLC-DETC assay for oxidized methyl linoleate.

Five replicates per day, three different days.

Methyl Linoleate Within Day (Peak Area)

Conc (uM) Day1 Day2 Day3 Between Day

10000 Ave 538113 559270 560980 552788

Stdev 5572 8799 12199 12737

1000 Ave 241431 335802 264894 280709

Stdev 21736 134629 9098 49133

750 Ave 154429 178414 180044 170963

Stdev 16009 7018 3399 14342

500 Ave 121257 122541 126650 123483

Stdev 7887 6810 1316 2817

250 Ave 14285 68231 68018 50178

Stdev 1533 3245 1615 31085

100 Ave 5424 5706 5877 5669

Stdev 171 312 722 229

Table 24. Reproducibility of HPLC-DETC assay for oxidized corn oil. Five replicates

per day, three different days.

Corn Oil Within Day (Peak Area)

Concentration (uM) Day1 Day2 Day3 Between Days

10000 Ave 475227 498878 467920 480675

Stdev 15989 10602 24328 16182

1000 Ave 66163 71271.2 80348 72594

Stdev 5665 824 5557 7184

750 Ave 54951 59653 59775 58126

Stdev 1511 1216 3471 2751

500 Ave 51291 43470 54547 49769

Stdev 1548 1177 2996 5693

250 Ave 6418 17312 17831 13854

Stdev 1629 1091 2069 6445

100 Ave 3899 19656 5032 9529

Stdev 451 21071 738 8788

91

Handling and Precautions:

In general, the DETC reagent and reaction mixtures are much more stable than the

HBr and NBP. However, use of HPLC introduces some additional complications.The

mobile phase (combination of water and acetonenitrile) developed for this assay is

sufficiently strong to elute epoxides up to C10 or possibly C12 but is inadequate to

move long chain fatty acids through the column in reasonable time, and it does not

dissolve triacylglycerols. Not eluting the epoxide complexes may be one reason for the

lower levels of epoxides detected in oils by this assay. The gradient program has been

modified to accommodate long chain epoxides and triacylglycerols, and this assay

needs to be retested using the new solvent combination.

4.4 NMR (Nuclear Magnetic Resonance) Assay

Use of NMR to measure epoxide groups in compounds is fast and convenient. It

can be preformed directly on extracts or oils and does not require any sample

modification other than dissolving in deuterated solvents. Certainly NMR is an

expensive instrument and is not available in every laboratory, and within institutions

NMR analyses are uaually available only on a fee basis and with restrictions on access

time. Because of these limitations, only three substances ( C10 1,2-epoxy-9-decene,

oxidized corn oil and oxidized methyl linoleate ) were evaluated by NMR.

92

Both 1H and

13C NMR were initially evaluated.

13C NMR more clearly

distinguished epoxide signals but sensitivity was substantially lower (Figure 21 A, B).

In 1H NMR spectra, the three hydrogens on the terminal epoxy ring present a triplet

specifically in the specific chemical shift range between 2.5~3.5 ppm, while in 13C

spectra, the two carbons of the epoxide give two lines in the range 45-55 ppm.These

values are consistent with those reported for other terminal epoxides (Gunstone and

Knothe 2010), and the specified peaks are not present in decene (Figure 21 C,D). For

both forms of NMR, the peak areas were proportional to the standard epoxide

concentration analyzed.

Detection limits and linearity of assay. The concentration response curve for 1H

NMR of 1,2-epoxy-9-decene is shown in Figure 22. As has been seen in all the assays

tested in this study, the concentration response was biphasic. There was a plateau range

from about 2 to 10 mM where the slope constantly varied. Below and above this range,

response was linear, although the slope was considerably steeper in the low

concentration range. This behavior is again consistent with all the other assays is fitting

a third order polynomial regression equation almost perfectly (R2=0.994) (Figure 22,

bottom). For this epoxide, the lowest concentration detectable and quantifiable by

NMR was 0.5 mM. Below this concentration, epoxide peaks could be seen but were

barely distinguishable from background noise.

93

Figure 21. 1H and

13C NMR spectra of 1,2-epoxy-9-decene. A and B: this study. C and

D. Decene spectra fromSpectral Database for Organic Compounds (Yamaji et al. 2013).

C

D

Epoxide peaks absent

B

A

1H NMR: 1-decene

STDB No. 2024HSP-02-010

13C NMR: 1-decene

SDBS No. 2024CDS-04-839

Epoxydecene

Epoxide peaks absent

9 8 7 6 5 4 3 2 1 0

10 9 8 7 6 5 4 3 2 1 0

94

Figure 22. 1H NMR peak area as a function of log epoxide concentration for

epoxydecene. Top: full concentration range. Bottom: lower concentration range.

y = 4.3799ln(x) + 35.485 R² = 0.9724

0

5

10

15

20

25

30

0.00 0.02 0.04 0.06 0.08 0.10

are

a u

nd

er

pe

ak

Epoxide concentration (M)

y = 5E+07x3 - 933764x2 + 5667.7x + 0.6146 R² = 0.994

0

2

4

6

8

10

12

14

16

0.000 0.002 0.004 0.006 0.008 0.010

are

a u

nd

er

peak

Epoxide concentration (M)

95

Figure 23. 1H NMR peak area as a function of log epoxide concentration for

epoxydecene. Top: full concentration range. Bottom: lower concentration range.

y = 8.9669x - 21.217 R² = 0.981

0

2

4

6

8

10

12

14

2 2.5 3 3.5 4

pe

ak a

rea

log [epoxide] (uM)

Epoxydecene

y = 10.085x - 25.025 R² = 0.9724

0

5

10

15

20

25

30

0 1 2 3 4 5 6

peak a

rea

log [epoxide] (M)

96

Application to oxidized lipids. Methyl linoleate and corn oil oxidized three days

at 40 C were analyzed by NMR over a range of oil concentrations. 1H NMR spectra are

shown in Figure 24. The three peaks within the epoxide chemical shift range 2.5~3.5

ppm were clearly present and reasonably strong. Plotting the integrated area under

these three peaks against oil concentration (note oil, not epoxide) yielded the response

curves shown in Figure 25.

The lowest detection limit for oxidized methyl linoleate we found is 0.25 mM oil

(lower than epoxydecene); the lowest detection limit we found for oxidized corn oil is

0.5 mM (same as epoxydecene). As with epoxydecene in 1H NMR and the epoxides in

general in all the assays, the concentration response was biphasic, increasingly rapidly

at low concentrations up to about 2 mM, then slowing at higher concentrations.

However, responses were non-linear in all concentration regions (Figure 25).

Since the response curves over the full concentration fit log relationships

reasonably well, the data was replotted based on log oil concentrations. As shown in

Figure 26, the concentration relationships to response became even more complex

rather than simplified to linear. Thus, in these oils, some complex factors other than oil

concentration are controlling what is detected by NMR. One possibility is viscosity of

the oil as the concentration increases. This phenomenon needs to be elucidated to

provide a valid basis for quantitating epoxides in oils by NMR.

97

Figure 24. 1H NMR spectra of oxidized methyl linoleate (top) and corn oil (bottom).

Red circles denote the epoxide peaks that were integrated to determine epoxide

response.

Corn oil 1

H NMR

Methyl linoleate 1H NMR

10 9 8 7 6 5 4 3 2 1 0 ppm

10 9 8 7 6 5 4 3 2 1 0 ppm

98

Figure 25. 1H NMR peak area as a function of oil concentration for oxidized methyl

linoleate and corn oil. Top: full concentration range. Bottom: lower concentration

range.

y = 2.7972ln(x) + 6.064 R² = 0.9066

y = 2.3685ln(x) + 4.023 R² = 0.9851

0

2

4

6

8

10

12

14

16

0 5 10 15 20

peak a

rea

concentration (mM)

y = 2.6756x2 + 0.4093x + 2.4189 R² = 0.9906

y = -5.796x2 + 12.281x - 2.6235 R² = 1

0

1

2

3

4

5

6

0 0.2 0.4 0.6 0.8 1 1.2

peak a

rea

Concentration (mM)

Methyl linoleate

Corn oil

Methyl linoleate

Corn oil

99

Figure 26. 1H NMR peak area as a function of log[oil] for oxidized methyl linoleate

and corn oil. Top. Full concentration range. Bottom: lower concentration range.

y = 39.719x3 - 314.26x2 + 829.61x - 727.33 R² = 1

y = -9.54x2 + 60.326x - 91.255 R² = 1

0

1

2

3

4

5

6

2.2 2.4 2.6 2.8 3 3.2

pe

ak a

rea

log[epoxide] (uM)

y = -10.116x3 + 98.566x2 - 304.86x + 306.54 R² = 0.9937

y = -1.4623x2 + 15.819x - 30.096 R² = 0.9964

0

2

4

6

8

10

12

14

16

18

2.0 2.5 3.0 3.5 4.0 4.5

peak a

rea

log concentration (uM)

Methyl linoleate

Corn oil

Methyl linoleate

Corn oil

100

Reproducbility of assay. Documentation of within day and between day

variability in NMR analyses is presented in Table 25 for epoxydecene, Table 26 for

methyl linoleate, and Table 27 for corn oil. Coefficients of variation for 1H NMR

analysis of epoxydecene and oxidized methyl linoleate and corn oil are listed in Table

28. Reproducibility of epoxydecene analysis by NMR was the best of all the assays

tested – about 1% within day and between days. Variability of the oils was slightly

higher – about 5% within day and slightly higher between days, but oth values were

well below the 10% limit usually applied to quantitative analyses. Examination of the

day to day epoxide levels shows that the variation was not random, but systematic,

decreasing on each successive day. Thus, the variation is due not to the NMR

instrumentation itself but to non-homogeneity of oils and to slow degradation or

continued transformation of the epoxides in the oil after the samples were oxidized and

then stored refrigerated.

Considering the high precision attainable with this assay and the sub-millimolar

sensitivity, NMR assays merit additional development.

101

Table 25. Reproducibility of NMR assay for C10 1,2-epoxy-9-decene. Five replicates

per day, three different days.

1,2-epoxy-9-decene Within Day Integrated Peak Areas

Conc. (M) Day 1 Day 2 Day 3 Between Days

0.1 Ave 27.7733 27.3067 26.5133 27.1980

Stdev 0.4771 1.7878 1.9634 0.6370

0.01 Ave 13.8633 14.4533 13.8633 14.0600

Stdev 1.6187 0.9816 1.6187 0.3406

0.0075 Ave 12.1000 11.7867 12.1000 11.9960

Stdev 0.3651 0.3885 0.3651 0.1809

0.005 Ave 11.5167 11.6267 11.5167 11.5530

Stdev 0.2730 0.3646 0.2730 0.0635

0.0025 Ave 10.0000 9.8800 10.0000 9.9600

Stdev 0.0600 0.0300 0.0600 0.0693

0.001 Ave 5.1133 5.2967 5.1133 5.1744

Stdev 0.1557 0.1850 0.1557 0.1058

0.00075 Ave 4.8667 4.9067 4.8667 4.8800

Stdev 0.1457 0.2001 0.1457 0.0231

0.0005 Ave 2.8033 2.9900 2.8033 2.8656

Stdev 0.5437 0.1153 0.5437 0.1078

102

Table 26. Reproducibility of NMR assay for epoxides in oxidized methyl linoleate.

Five replicates per day, three different days.

Methyl linoleate Within Day Integated Peak Area

Conc. (M) Day1 Day2 Day3 Between Days

0.02 Ave 14.1433 14.1067 13.1067 13.7860

Stdev 0.5181 0.2003 0.2003 0.5882

0.002 Ave 10.4833 11.1000 10.1567 10.5800

Stdev 0.8223 0.5292 0.0404 0.4790

0.001 Ave 5.9433 5.5067 5.2000 5.5500

Stdev 0.1150 0.4727 0.4678 0.3736

0.0008 Ave 4.2367 4.0367 4.0033 4.0922

Stdev 0.2237 0.8370 0.2409 0.1262

0.0005 Ave 3.2600 3.7100 3.3233 3.4311

Stdev 0.2869 0.9305 0.1701 0.2436

0.0003 Ave 2.8400 2.5333 2.5533 2.6422

Stdev 0.3208 0.3412 0.1914 0.1716

103

Table 27. Reproducibility of NMR assay for epoxides in oxidized corn oil. Five

replicates per day, three different days.

Corn Oil Within Day Integrated Peak Areas

Conc. (M) Day1 Day2 Day3 Between Day

0.02 Ave 11.0033 10.8067 10.8133 10.8740

Stdev 0.0473 0.1716 0.1888 0.1117

0.002 Ave 6.6767 6.2800 6.2400 6.3989

Stdev 0.2098 0.2390 0.0954 0.2414

0.001 Ave 4.5100 3.6633 3.4100 3.8611

Stdev 0.5507 0.5082 0.4597 0.5761

0.00075 Ave 3.7533 3.3200 2.9067 3.3267

Stdev 0.3109 0.3961 0.0306 0.4234

0.0005 Ave 2.0333 2.2333 1.9367 2.0678

Stdev 0.3570 0.2401 0.2139 0.1513

104

Table 28. Coefficients of variation for NMR assay of epoxydecene and epoxides in

oxidized lipids. Averaged over all concentrations and days.

average COV(%)

Within day Between days

C10 1.01 1.72

Corn oil 5.32 7.95

ML 4.52 5.36

Handling and Precautions. 1H NMR analyses specifically detected proton

interactions in molecules, including any water present. Thus, it is crucial to exclude

water from all test samples and solvents. In addition, also to avoid interference from

extraneous protons, deuterated solvents must be used, e.g. deuterated

chloroform (CDCl3) for lipids. These solvents do degrade with time so it is prudent to

purchase only small quantities of these solvents at a time and to not store these reagents

for long periods.

105

5. SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS

Since the importantce of lipid oxidation was recognized more than 50 years ago,

epoxides have been overlooked as lipid oxidation products. Although epoxides are

recognized as products, they are seldom analyzed or considered when examining

reaction processes, perhaps because they form and react rapidly so are seldom detected

among mixed products, because they do not produce known off-odors or flavors, or

because there are few standard procedures or guides for analyses of epoxides. The main

purpose of this thesis research was to evaluate four available epoxide assays to

determine detection ranges, accuracy, reproducibility, and special handling issues in

order to provide some guidelines and protocols for scientists for selecting and using

appropriate methods when quantitating epoxides. Our main focus was on accurate

quantitation rather than relative comparisons since we need to use the results in

developing mass balances of alternate reaction pathways in lipid oxidation.

A major limitation to accurate quantitation that arose in the HBr and NBP assays

was marked differences in reactivity that depended on epoxide structure. In these two

assays, epoxide reactivity increased with epoxide chain length. One potential

explanation for this enhancement is that the longer alkyl groups increase polarity of the

oxirane ring more, and hence facilitate attack by nucleophiles such as HBr and NBP.

Another possibility, at least for NBP, is that the NBP-epoxy adduct structure alters the

106

extinction coefficient and optical properties that are the basis for detection. Whatever

the explanation, non-constant reaction with epoxides is not an acceptable basis for

quantitative analyses of materials that have different or unknown epoxide structures, as

occurs with different oils and lipid extracts of materials.

Additional limitations that we noted in some assays was sensitivity to oxygen and

volatility/reactivity of epoxides. To maximize reagent stability and minimize side

reactions, we found it prudent to flush all samples, solutions, and sample headspaces

with argon and to minimize exposure to light as well. Samples, especially epoxide

standards, must be handled, transferred, and capped rapidly to avoid degradation and

loss. Strong odors in the laboratory are certain signs of epoxide loss.

Detection range is critically important when choosing a suitable assay. Detection

ranges and sensitivity vary with the assay. If you choose an assay with a detection range

that does not match your samples, the test results may well be erroneous, misleading,

and unreliable. The detection ranges for assays evalutated in this study are summarized

in Table 29. Perhaps because of limitations in useful stable HBr concentrations, the

HBr titration assay has a narrow detection range of only 0.1 M down to 7.5 mM epoxide.

This assay was originally developed to analyze epoxides in modified fuels and greases

where epoxide concentrations are high, and is probably borderline in being ablt to

detect epoxides in early stages of lipid oxidation. If this was applied to analyze lipid

107

Table 29. Summary of detection ranges and important characteristics of HBr, NBP,

DETC, and NMR epoxides

HBr Assay

Limits of detection 7.5 mM to 0.1 M (1M for epoxybutane)

Linear detection range non-linear response curves at all concentrations

Epoxide structure effects marked, reaction increases with R chain length

Oxygen sensitive? results lower in air than under Ar

Reproducibility (%) Average within day Ave between day

Standards 2-3% 3-9%

Oils 4-6% 8-13%

Handling issues HBr volatile and degrades rapidly, frequent

restandardization is necessary

Endpoint difficult to determine visually

Reaction less than stoichiometric, even under Ar

Sparging with inert gas (Ar) increases yields --

inhibits Br- oxidation and epoxide side reactions

Advantages fast and relatively simple, standard equipment

Recommendations for use use only for samples with high epoxides.

can compare only samples with similar epoxide

structures

p-Nitrobenzyl Pyridine (NBP) Assay

Limits of detection C4: 10 M - 1M; C6: 1M to 100 mM;

C10: 1M – 50 mM

Linear detection range 1 M to 1 M for epoxybutane

1-10 M for epoxyhexene

polynomial at all concentrations for epoxydecene

Epoxide structure effects pronounced, reaction increases with R chain length

Oxygen sensitive? Yes, but with tube capped all the time it works

Reproducibility (%) Average within day Ave between day

Standards 1-2% 2-3%

Oils 3-7% 4-12%

Handling issues reactions affected by unknown factors other than

epoxide concentration

Advantages high sensitivity and reproducility

Recommendations for use gives relative rather than absolute qunatitation

can compare only samples with similar epoxide

structures

108

Diethyldithiocarbamate (DETC)-HPLC Assay

Limits of detection 2 M, C6 and C10; 20 M C4 epoxides

upper limit ~10 mM (solubility limit of DETC)

Linear detection range 2 M~100 M epoxides

appears to have a second linear range with lower slope

at high epoxide concentrations

Epoxide structure effects effects minimal mid chain length and above

Oxygen sensitive? No effect noted

Reproducibility (%) Average within day Ave between day

Standards 3% 5-7%

Oils 11-18% 16-28%

Handling issues current procedures designed for monomer epoxides,

HPLC gradient modifications required for epoxides on

triacylglycerols and large molecules, in extracts

Advantages fast, easy, sensitive

provides information about individual epoxide species

adaptable to microplate solution assays

Recommendations for use Useful for research and quality control with gradient

adjustment

Merits detailed development and further testing

NMR Assay

Limits of detection 0.5 mM

Linear detection range 0.5 mM – 100 mM

Epoxide structure effects Apparently minimal differences when samples are

not volatile

Oxygen sensitive? No, except for stability of the sample itself.

Reproducibility (%) Average within day Ave between day

Standards 1% 2%

Oils 5% 5-8%

Handling issues Requires dry, deuterated solvents

Potential interferences in mixed extracts unknown

Advantages Does not require derivatization or reaction

Performed directly on oils, extracts, pure samples

Provides both absolute quantitation and qualitative

information about epoxide structure

Recommendations for use Merits detailed development and further testing

109

oxidation in foods or biological tissues in the past, it most likely resulted in conclusions

that epoxides were not present.

Based on accuracy of chemistry, linearity of response, sensitivity, lack of epoxide

structure effects on reaction response or detection, and ability to obtain absolute

quantitation, only two assays – DETC and NMR – are considered to be candidates for

further development as routine analyses of lipid oxidation.

The DETC-HPLC assays seem to be the most promising method because of its

sensitivity, linear response, and ability to separate epoxides of different structures.

Combining this with mass spectrometry will create a very powerful tool for research on

oxidizing lipids. The main limitation currently is the eluting solvent and gradient

program, which was developed for monomer epoxides. When the solvent program

recently developed in this laboratory is fully tested, the DETC-HPLC assay will be a

very powerful tool indeed for analysis of oxidized lipids.

NMR is a very intriguing approach to analyzing lipid epoxides. Minimal handling

is required – just dissolving sample or oil in deuterated solvents – and actual analysis is

straightforward if NMR instrumentation is available. Samples can be analyzed in about

10 minutes on a 400 MHz NMR, with sufficient resolution and sensitivity. NMR also

offers the potential to detect other oxidation products in the same same since

oxygenated groups are shifted substantially downfield and are separated from each

110

other. values for each are substantially different. Responses were linear and sensitive

enough (about 0.5 mM as a lowe limit) to detect epoxides in many oil samples from the

laboratory. It may be possible to increase sensitivity and resolution by switching to

higher frequency NMRs, but this is probably more for research due to its high

associated user fees. Some attention will need to be given to effects of solvent and water

traces when extracts rather than refined oils are analyzed.

111

6. FUTURE WORK

To our knowledge, assays for lipid epoxides have not been evaluated previously, so

this study begins the process. All of the assays still need work to make them rugged and

dependable, and hopefully also more sensitive. As noted in the recommendations, the

HBr and NBP assays are not appropriate for analyses of epoxides with mixed or

unknown structures. However, the DETC and NMR assays offer potential for sensitive

detection independent of epoxide structure, so merit considerable effort to optimize.

Modifications of the HPLC-DETC assay. Two major adaptations are needed

here. First, there is the question of whether epoxides in oils exist as monomer scission

products or are also retained on the triacylglycerols backbone, i.e. as core epoxides. In

its current form, the elution gradient probably grossly underestimates epoxide levels of

long chain fatty acids and triacylglycerols. This can be remedied by a) extracting the

monomer epoxides with methanol and analyzing without interference from larger

molecules, and b) adapting the solvent gradient to handle larger molecules such as

triacylglycerols. A solvent gradient has bee modified for elution of triacylglycerols and

this needs to be tested with epoxidized oils. In addition, stability of epoxides in lipid

solvents must be tested. We have some evidence from extraction studies that epoxides

degrade rapidly in organic solvents.

112

Questions have been raised about the specificity of the DETC assay. Our

preliminary analyses found no reaction with other lipid oxidation products such as

hydroperoxides and carbonyls, but this needs to be evaluated more systematically.

In addition, to make maximum use of the HPLC separation of individual epoxide

adducts, LC needs to be interfaced with mass spectrometry detection to identify

specific epoxide structures. This will contribute greatly to elucidating lipid oxidation

mechanisms. MS will also verify that peaks detected are indeed from epoxides and not

other degradation pathways or adventitious transformations. This information will

allow exclusion of non-targeted peaks from integration, hence increasing sensitivity

and improving accuracy of this assay.

Modifications of the NMR assay. Samples in this study were analyzed using a

400 MHz NMR spectrometer that provides middle of the road sensitivity. Since the end

goal is to proved an assay that can be used at multiple levels, e.g. in quality control

laboratories in industry as well as in academic research with materials ranging from

biological tissues with barely trace levels of epoxides to foods in very early stages of

oxidation to foods in later stages of degradation, it is imperative to determine what level

of NMR is required to provide the sensitivity needed for each application. Thus, studies

need to be conducted comparing sensitivity of table top NMRs to various frequency

NMRs, e.g. 200, 400, 600, 800 MHz spectrometers. These range from rather crude

113

analyses to high sensitivity research instruments and will reveal which level is needed

for different applications. In addition, pure material were used in this study, but lipid

extracts are the samples analyzed in most routine quality control as well as basic

research. We noted problems with interference from water and solvent protons. Both

may be a significant problem when analyzing extracts. Detailed investigation of the

effects of solvent and water traces as well as multiple components on NMR analyses

needs to be undertaken.

114

7. REFERENCES

Acott, B.and Beckwith, A. L. J. (1964) “Reactions of alkoxy radicals. IV.

Intramolecular hydrogen-atom transfer in the presence of cupric ion: A novel

directive effect” Austral J. Chem 17(12), 1342-1353

Aerts, H. A. J. and Jacobs, P.A. (2004). “Epoxide yield determination of oils and fatty

acid methyl esters using 1H NMR.” Journal of the American Oil Chemists’ Society

81 (9), 841–846.

Agarwal, S. C., Van Duuren, B. L. and Kneip, T. J. (1979). “Detection of epoxides with

4-(p-nitrobenzyl) pyridine.” Bulletin of Environmental Contamination And

Toxicology 23 (6), 825–829.

American Oil Chemistry Society Official Methods (1997). Oxirane Oxygen. Method

Cd 9-57.

Anderson, H. C. (1962), Thermal degradation of epoxide polymers. J. Appl. Polym.

Sci., 6, 484-488.

Ansari, M.H. and Ahmad, M. (1986). “Acacia suma seed oil - characterization of a

HBr-reactive acid.” Fette, Seifen, Anstrichmittel 88 (10), 402–403.

Baumann, K., Hogenauer, K., Schulz, G. and Steck, A. (2002). “NMR study on 9-S and

9-R oxirane derivatives of ascomycin.” Magnetic Resonance in Chemistry 40 (7),

443–448.

Benson, A. M., Barretto, P. B. (1985). “Effects of disulfiram , diethyldithiocarbamate,

bisethylxanthogen , and benzyl isothiocyanate on glutathione transferase activities

in mouse organs.” Cancer Research 45, 4219–4223.

Bush, D. M. and Kozumbo, J.W. (1983). “Metabolism of polycyclic aromatic

hydrocarbon derivatives to ultimate carcinogens during lipid peroxidation.”

Science 221 (4605), 77-79.

Chan, H. W. S., Prescott, F. A. A., and Swoboda, P. A. T. (1976) Thermal

decomposition of individual positional isomers of methyl linoleate hydroperoxide:

Evidence of carbon-oxygen bond scission. J. Am. Oil Chem. Soc 53, 572-576.

Chen, H.J., Gonzalez, F. J., Shou, M. and Chung, F. L. (1998).

“2,3-Epoxy-4-hydroxynonanal, a potential lipid peroxidation product for etheno

115

adduct formation, is not a substrate of human epoxide hydrolase.” Carcinogenesis

19 (5), 939–43.

Claxson, A.W.D., Hawkes, G. E., Richardson, D. P., Naughton, D. P., Haywood, R.M.,

Chander, C. L., Atherton, M., Lynch, E. J., Grootveld, M. C. (1994). “Generation

of lipid peroxidation products in culinary oils and fats during episodes of thermal

stressing: a high field 1H NMR study.” FEBS letters 355 (1), 81–90.

Clayden, J., Greeves, N. and Warren S. ed.(2006) Organic Chemistry, Oxford

University Press, Oxford.

Cravero, R. M., Labadie G. R. and Sierra M. G. (2000). “Using empirical rules from 13

C

NMR analysis to determine the stereochemistry of the epoxide located at the

5,6-position of decalinic systems.” Molecules 5 (3), 323–324.

Derawi, D. and Salimon, J. (2010). “Optimization of epoxidation of palm olein by using

performic acid.” E-Journal of Chemistry 7 (4), 1440–1448.

Dupard-Julien, C. L., Kandlakunta, B., and Uppu, R. M. (2007). “Determination of

epoxides by high-performance liquid chromatography following derivatization

with N,N-diethyldithiocarbamate.” Analytical and Bioanalytical Chemistry 387

(3), 1027–32.

Durbetak, A. J. (1958). “Determination of oxirane oxygen in salts of epoxy acids and in

the presence of amines.” Anaytical Chemistry 30 (12), 2024-2025.

Earle, F. R. (1970). “Epoxy oils from plant seeds.” Journal of the American Oil

Chemists' Society 47 (12), 510–513.

Ehsan, S.D. (2010). “Optimization on Epoxidation of Palm Olein by Using Performic

Acid.” E-Journal of Chemistry 7(4):1440–1448.

Ergozhin, E. E., Bektenov, N. A. and Chopabaeva, N. N. (2004). “NMR study of the

structure of polymers based on 2,3-epoxypropyl methacrylate.” Russian Journal

of Applied Chemistry 77 (5), 813–816.

Fankhauser-Noti, A., Fiselier, K., Biedermann-Brem, S. and Grob, K. (2006).

“Assessment of epoxidized soy bean oil (ESBO) migrating into foods: comparison

with ESBO-like epoxy fatty acids in our normal diet.” Food and Chemical

Toxicology 44 (8), 1279–86.

116

Fennema, O. R., ed. (1996). Food Chemistry, 3rd Edition, Macel Dekker Inc. New

York.

Frankel, E. N., ed. (1998). Lipid Oxidation , The Oily Press, Scotland.

Frankel, E. N., Neff, W. E., Selke, E., and Weisleder, D. (1982) Photosensitized

oxidation of methyl linoleate: Secondary and volatile thermal decomposition

products. Lipids 17, 11-18.

Gardner, H., .W., Weisleder, D., and Kleiman, R. (1978) Formation of

trans-12,13-epoxy-9-hydroperoxy-trans-10-octadecadienoic acid catalyzed by

either a soybean extract or cysteine-fecl3. Lipids 13, 246-252.

Gardner, H. W. and Kleiman, R. (1981) Degradation of linoleic acid hydroperoxides by

a cysteine-FeCl3 catalyst as a model for similar biochemical reactions. II.

Specificity in formation of fatty acid epoxides. Biochim. Biophys. Acta 665,

113-125.

Giuffrida, F., Destaillats, F., Robert, F., Skibsted, L. H. and Dionisi, F. (2004).

“Formation and hydrolysis of triacylglycerol and sterols epoxides: role of

unsaturated triacylglycerol peroxyl radicals.” Free Radical Biology & Medicine

37 (1),104–14.

Guillen, M. D. and Goicoechea, E. (2009). “Oxidation of corn oil at room temperature:

primary and secondary oxidation products and determination of their

concentration in the oil liquid matrix from 1H nuclear magnetic resonance data.”

Food Chemistry 116 (1), 183–192.

Guillén, M. D. and Ruiz, A. (2006). “Study by means of 1H nuclear magnetic resonance

of the oxidation process undergone by edible oils of different natures submitted to

microwave action.” Food Chemistry 96 (4), 665–674.

Guillén, M. D. and Ruiz, A. (2005). “Monitoring the oxidation of unsaturated oils and

formation of oxygenated aldehydes by proton NMR.” European Journal of Lipid

Science and Technology 107 (1), 36–47.

Guillén, M. D. and Ruiz, A. (2004). “Study of the oxidative stability of salted and

unsalted salmon fillets by 1H nuclear magnetic resonance.” Food Chemistry 86 (2),

297–304.

117

Gunstone, F. and Knothe, G. (2010) NMR spectroscopy of fatty acids and their

derivatives., Lipid Library, http://lipidlibrary.aocs.org/nmr/nmr.html, accessed

April 14, 2013.

Guo, R., Ma, C., Sun, S. and Ma, Y. (2010). “Kinetic study on oxirane cleavage of

epoxidized palm oil.” Journal of the American Oil Chemists’ Society 88 (4),

517–521.

Hammock, L. G., Hammock, B. D., and Casida, J. E. (1974). “Detection and analysis of

epoxides with 4-(p-nitrobenzyl)-pyridine.” Bulletin of Environmental

Contamination and Toxicology 12 (6), 759–764.

Haynes, S.C. and Vonwiller, R.K. (1990) Iron(III) and copper(II) catalyzed

transformation of fatty acid hydroperoxides: efficient generation of peroxy

radicals with Cu(II) trifluoromethane sulfonate. J. Chem. Soc. Chem. Commun.,

1102-1104.

Hemminki, K. (1979). “Correlation of mutagenicity and 4-(p-nitrobenzyl)-pyridine

alkylation by epoxides” Toxicology Letters 4, 103–106.

Hemminki, K., Falck, K., and Vainio, H. (1980). “Comparison of alkylation rates and

mutagenicity of directly acting industrial and laboratory chemicals.” Archives of

Toxicology 46 (3-4), 277–285.

Hemström, P. and Irgum, K. (2006). "Hydrophilic interaction chromatography."

Journal of Separation Science 29 (12), 1787-1821.

Husain, S. R., Ahmad, M. S., Ahmad, Jr. F., Ahmad, M., and Osman, S. M. (1989).

“Hibiscus mutabilis seed oil - characterization of HBr-Reactive Acids.” Fat

Science and Technology 91 (4), 167–168.

Kamal-Eldin, A., Marquez-Ruiz G., Dobarganes, C. and Appelqvist, L. A. (1997).

“Characterisation of aldehydic acids in used and unused frying oils.” Journal of

Chromatography 776 (2), 245–254.

Kan, C. C. (2006). “Lipid” In Food Chemistry, 1st Edition, Lun, X. ed, New Wun Ching

Developmental Publishing Co., Ltd., Taipei, pp.114-165.

Kannan, R., Subbaram, M. R., and Achaya, K. T. (1974). “NMR studies of some

oxygenated. halogenated, and sulphur-containing fatty acids and their derivatives.”

Fette, Seifen, Anstrichmeittel 76, 344–350.

118

Karel, M. (1980) Lipid oxidation, secondary reactions, and water activity of foods. In

Autoxidation in food and biological systems. Simic, M. G., & Karel, M., eds, New

York: Plenum Press, pp. 191-206

Kim, J. H. and Thomas, J. J. (1992). “Use of 4-(nitrobenzyl)pyridine (4-NBP) to test

mutagenic potential of slow-reacting epoxides, their corresponding olefins, and

other alkylating agents.” Bulletin of Environmental Contamination and

Toxicology 49 (6), 879–85.

Kochi, J. K. (1962) Chemistry of alkoxyl radicals. J. Am. Chem. Soc. 84, 1193-1197.

Kochi, J. K. (1973) Oxygen radicals. In Free Radicals, Vol. 2 (Kochi, J. K., ed), John

Wiley, New York, pp. 665-710.

Kuo, T. M. and Gardner, H. W. ed. (2002) Lipid Biotechnology, Marcel Dekker Inc,

New York, NY

Labuza, T. P. (1972). “Kinetics of lipid oxidation in foods.” CRC Critical Reviews in

Food Technology 2 (3), 355–405.

Lederer, M. O. (1996). “Reactivity of lysine moieties toward γ-hydroxy-α,β

-unsaturated epoxides: a model study on protein-lipid oxidation product

interaction." Journal of Agriultural and Food Chem. 44 (9), 2531–2537.

Lee, C. S., Ooi, T. L., and Chuah, C. H. (2009). “The effect of reaction temperature on

retaining oxirane oxygen contents in the synthesis of epoxidized diethanolamides.”

American Journal of Applied Science 6 (1), 72–77.

Lee, S. H., Oe, T., and Blai, I. A. (2002). “4,5-epoxy-2(E)-decenal-induced formation

of 1,N-etheno-2'-deoxyadenosine and 1,N-etheno-2'-deoxyguanosine adducts.”

Chemical Research in Toxicology 15 (3), 300–304.

Lee, S. H., Oe, T., and Blair, I. A. (2001). “Vitamin C-induced decomposition of lipid

hydroperoxides to endogenous genotoxins.” Science 292 (5524), 2083–6.

Marcel, S. F., Jie, L. K., Lam, N. W., Ho, C. M. and Lau, M. L. (2003). “Epoxidation of

a conjugated linoleic acid isomer.” European Journal of Lipid Science and

Technology 105 (8), 391–396.

119

Marmesat, S., Velasco J., and Dobarganes, M. C. (2008). “Quantitative determination

of epoxy acids, keto acids and hydroxy acids formed in fats and oils at frying

temperatures.” Journal of Chromatography A 1211 (1-2), 129–34.

Meshram, P. D., Puri, R. G. and Patil, H. V. (2011). “Epoxidation of wild safflower

(carthamus oxyacantha ) oil with peroxy acid in presence of strongly acidic cation

exchange resin IR-122 as catalyst.” International Journal of ChemTech Research

3 (3), 1152–1163.

Moreno, M., Olivares, D. M., Amezquita Lopez, F. J., Martines, V. P., and Reig, F. B.

(1999). “Study of the formation of carbonyl compounds in edible oils and fats by 1H-NMR and FTIR.” Journal of Molecular Structure 483, 557–561.

Munger, D., Sternson, L. A., Repta, A. J., and Higuchi, T. (1977). “High-performance

liquid chromatographic analysis of dianhydrogalactitol in plasma by

derivatization with sodium diethyldithiocarbamate.” Journal of Chromatography

143 (4), 375–82.

Nawar, W. W. (1969) "Thermal degradation of lipids" J. Agr. Food Chem 17, 18-21.

Nawar, W. W. (1986) Lipids. In Fennema, Food Chemistry pp. 225-320. New York:

Academic.

Nelis, H. J., Alry, S. C. and Sinsheimer, J. E. (1982). “Comparison of the alkylation of

nicotinamide and 4-(p-nitrobenzyl)pyridine for the determination of aliphatic

epoxide.” Analytical Chemistry 54, 213–216.

Nelis, H. J., and Sinsheimer, J. E. (1981). “A sensitive fluorimetric procedure for the

determination of aliphatic epoxides under physiological conditions.” Analytical

Biochemistry 115 (1), 151–7.

Nielsen, H. K., Löliger, J. and Hurrell, R. F. (1985). “Reactions of proteins with

oxidizing lipids.” British Journal of Nutrition 53 (1), 61-73.

Piazza, G. J., Nuñez, A., and Foglia, T. A. (2003). “Epoxidation of fatty acids , fatty

methyl esters , and alkenes by immobilized oat seed peroxygenase.” Journal of

Molecular Catalysis 21, 143–151.

Polozov, G. I., Tishchenko, I. G., Suboch, V. P. (1986). “Mass spectra and 13

C NMR

spectra of the adducts of epoxyenamines with sulfene.” Chemistry of Heterocyclic

Compounds 22 (2), 217–221.

120

Rather, L. J., Knapp, B., Haehnel, W. and Fuchs. G. (2010). “Coenzyme A-dependent

aerobic metabolism of benzoate via epoxide formation.” Journal of Biological

Chemistry 285 (27), 20615–24.

Rebsdat, S., Mayer, D., ed (2001) “Ethylene Oxide” In Ullmann's Encyclopedia of

Industrial Chemistry. Wiley-VCH, Weinheim, Germany.

Schaich, K. M. (1980). “Free radical initiation in proteins and amino acids by ionizing

and ultraviolet radiations and lipid oxidation. Part II. Ultraviolet radiation and

photolysis.” Critical Reviews in Food Science and Nutrition 13 (2), 131–159.

Schaich, K. M. (1992). “Metals and lipid oxidation. Contemporary issues.” Lipids 27

(3), 209–218.

Schaich, K. M. (2005). “Lipid oxidation: theoretical aspects.” In Bailey’s Industrial Fat

and Oil Products, 6th

Edition, Vol.1, Shahidi, F., John Wiley, ed, New York, pp.

269-355.

Schaich, K. M. (2008). “Co-oxidation of proteins by oxidizing lipids. In Lipid

Oxidation Pathways, Vol. 2., Kamal-Eldin, A. and Min, D. ed, AOCS Press,

Champaign, IL, pp. 181–272.

Schaich, K. M. (2012). “Lipid oxidation.” In Biochemistry of Foods. Eskin, N.A.M. and

Shahidi, F. ed, Academic Press, New York, pp. 419–478.

Schaich, K. M. (2012). “Thinking outside the classical chain reaction box of lipid

oxidation.” Lipid Technology 24 (3), 55–58.

Sevanian, A., Mead, J. F. and Stein, R. A. (1979). “Epoxides as products of lipid

autoxidation in rat lungs.” Lipids 14 (7), 634–43.

Shahidi, F. and Wanasundara, U. N. (1996). “Methods for evaluation of the oxidative

stability of lipid-containing food.” Food Science and Technology 2 (2), 73–81.

Solomons, T.W.G. (2004) “Epoxides” In Organic Chemistry, Brennan, D., ed, John

Wiley and Sons. Inc, Danvers, CT, pp. 516-519.

Spiteller, G. (2003). “Are lipid peroxidation processes induced by changes in the cell

wall structure and how are these processes connected with diseases” Medical

Hypotheses 60 (1), 69–83.

121

Thomas, J. J., Kim, J. H. and Mauro, D. M. (1992). “4-(4-Nitrobenzyl)pyridine tests for

alkylating agents following chemical oxidative activation.” Archives of

Environmental Contamination and Toxicology 22 (2), 219–227.

Tung, H.C., Kang, C. and Sawyer, D. T. (1992) “Nature of reactive intermediate from

the iron-induced activation of hydrogen peroxide agents for the ketonization of

methylenic carbons, the monooxygenation of hydrocarbons and the dioxygenation

of arylolefins” J. Amer. Chem. Soc 114, 3445-3455

Van Damme, F., and Oomens, A.C. (1995). “Determination of residual free epoxide in

polyether polyols by derivatization with diethylammonium

N,N-diethyldithiocarbamate and liquid chromatography.” Journal of

Chromatography A 696 (1), 41–47.

Wu, G. S., Stein, R. A., Mead, J. F. (1977). “Autoxidation of fatty acid monolayers

adsorbed on silica gel.” LIPIDS 12 (11), 971-978.

Yamaji, T., Saito, T., Hayamizu, K., Yanagisawa, M. and Yamamoto, O. (2013) 13

C

NMR of 9,10-epoxy-1-decene- Spectral Database for Organic Compounds, SDBS.

National Institute of Advanced Industrial Science and Technology (AIST) Japan,

http://sdbs.riodb.aist.go.jp/sdbs/cgi-bin/direct_frame_top.cgi, accessed April 12,

2013.

122

Curriculum Vitae

Chen-Hsiang Liao

Education

Rutgers University, New Brunswick, NJ May, 2013

M.S. in Food Science

National Taiwan University Jun, 2008

B.S. in Agriculture Chemistry

Research

Laboratory of Lipid Chemistry, Rutgers University, NJ Jan, 2011 to Jan, 2013

Project member

Laboratory of agriculture chemistry, Jun, 2007 to Jun, 2008

National Taiwan University, Taipei, Taiwan

Independent project leader

Experience

Graduate student assistant, Rutgers University, NJ Jan, 2011 to Jan, 2013

Member of GSA product development team Jan, 2011 to Jan, 2012

Rutgers University, NJ